Effects of conservation tillage practices on ammonia emissions from Loess Plateau rain-fed winter wheat fields

Atmospheric Environment 104 (2015) 59e68

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Effects of conservation tillage practices on ammonia emissions from Loess Plateau rain-fed winter wheat fields Yang Yang a, Chunju Zhou b, Na Li a, Kun Han c, Yuan Meng a, Xiaoxiao Tian a, Linquan Wang a, * a b c

College of Resources and Environment, Northwest A&F University, Yangling, Shaanxi Province 712100, China College of Life Science, Northwest A&F University, Yangling, Shaanxi Province 712100, China College of Agronomic Sciences, Shandong Agricultural University, Tai'an, Shandong Province 271018, China

h i g h l i g h t s
Ammonia emissions were measured using a vented chamber method.
Conservation tillage practices significantly reduce ammonia emissions.
Deep-band application of nitrogen fertilizer reduces ammonia emissions.
Film and stalk mulches can both reduce ammonia volatilization.
Ammonia fluxes are strongly dependent on soil ammonium, moisture, and temperature.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 March 2014 Received in revised form 19 November 2014 Accepted 4 January 2015 Available online 5 January 2015

Ammonia emissions from agricultural activities contribute to air pollution. For the rain-fed winter wheat system in the Loess Plateau there is a lack of information about ammonia emissions. Current study aimed to provide field data on ammonia emissions affected by conservation tillage practices and nitrogen applications. A two-year field experiment was conducted during 2011e2013 wheat growing seasons followed a split-plot design. Main plots consisted of one conventional tillage (CT, as the control) and five conservation tillage systems, i.e., stalk mulching (SM), film mulching (FM), ridge tillage (RT), ridge tillage with film mulch on the ridge (RTfm), and ridge tillage with film mulch on the ridge and stalk mulch in the furrow (RTfmsm); while subplots consisted of two nitrogen application rates, i.e., 0 and 180 kg N ha1. Ammonia emissions were measured using an acid trapping method with vented chambers. Results showed ammonia fluxes peaked during the first 10 days after fertilization. On average, nitrogen application increased ammonia emissions by 26.5% (1.31 kg N ha1) compared with treatments without nitrogen application (P < 0.05). Ammonia fluxes were strongly dependent on soil ammonium, moisture, and temperature. Tillage systems had significant effects on ammonia emissions. On average, conservation tillage practices reduced ammonia emissions by 7.7% (0.46 kg N ha1) compared with conventional tillage (P < 0.05), with FM most effective. Deep-band application of nitrogen fertilizer, stalk mulches, and film mulches were responsible for reductions in ammonia emissions from nitrogen fertilization in conservation tillage systems, thus they were recommended to reduce ammonia emissions from winter wheat production regions in the southern Loess Plateau. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Ammonia volatilization Acid trapping method Soil moisture Nitrogen fertilizer

1. Introduction Ammonia is the most abundant alkaline gas in the atmosphere,

* Corresponding author. E-mail addresses: [email protected] (Y. Yang), linquanw@nwsuaf. edu.cn (L. Wang). http://dx.doi.org/10.1016/j.atmosenv.2015.01.007 1352-2310/© 2015 Elsevier Ltd. All rights reserved.

with agriculture as the largest source (Behera et al., 2013; MartinezLagos et al., 2013). Nitrogen fertilization is a key driving force in the biogeochemical cycle of atmospheric ammonia. Over the last few decades, ammonia emission has been increasing on a global scale, with adverse effects on global climate change, environment, and public health (Behera et al., 2013). For example, deposited ammonia increases nitrification and denitrification rates and then increases

60

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

emissions of nitrous oxide, a potent greenhouse gas (Ferm, 1998). Ammonia reacts with acidic substances in the atmosphere to form ammonium salts which result in the formation of atmospheric particulate matter (e.g., PM2.5), haze pollution, and visibility degradation, posing a substantial threat to human health as well as traffic (Gong et al., 2013a). Ammonia emissions from agricultural activities are notable in China. For example, ammonia emission is 1.29e4.68 kg N ha1 (0.3e1.9% of applied N) from winter wheat fields in the southern Loess Plateau of China (Shangguan et al., 2012). Over the past decade, China has suffered serious haze pollutions (Ma et al., 2012a), partially due to the increasing atmospheric ammonia. Consequently, it is time for China to take actions to reduce ammonia emission and deposition in agricultural systems, and

alleviate their adverse impacts locally and globally (Liu et al., 2013). Winter wheat is one of the most important cereal crops in the southern Loess Plateau of China (Fig. 1a), with urea as the predominant nitrogen fertilizer, which is usually applied in a single application before sowing wheat (Roelcke et al., 1996). However, urea has the potential for substantial losses of ammonia (SanzCobena et al., 2011). Moreover, the high pH of calcareous soils in Loess Plateau results in great ammonia losses, with maximum losses reaching 50% of applied N (Roelcke et al., 1996). Therefore, strategies to reduce ammonia emissions from winter wheat fields in this region are critical for the nitrogen emission and deposition reduction in China. Conservation tillage practices have been recommended to maintain crop yield, leading to variations in ammonia emissions, which are sensitive to surface mulching

Fig. 1. Location of field site (a), plot layout (b), and scheme of tillage systems (c). The winter wheat production region is shown as the shaded area. Main plots (tillage systems) were split into two nitrogen fertilizer subplots.

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

(Shangguan et al., 2012) and fertilizer placement (Webb et al., 2014). Previous findings showed ridge tillage, a conservation tillage practice, can reduce ammonia emissions (Shangguan et al., 2012). However, little information is available regarding the conservation tillage effect simultaneously with nitrogen fertilization effect on ammonia emission in this region. The objectives of this study were to i) assess the effects of conservation tillage practices and nitrogen fertilization on ammonia emissions from winter wheat fields using vented chambers, ii) elucidate the mechanisms and provide a scientific basis for developing effective control strategies for ammonia emissions. 2. Materials and methods 2.1. Site description A two-year field experiment was conducted during 2011e2012 (Oct. 8, 2011 to Jun. 7, 2012) and 2012e2013 (Oct. 6, 2012 to May 27, 2013) wheat growing seasons. The study site (34170 3500 N, 108 0401200 E; 520 m ASL) was located in Yangling in the southern Loess Plateau, China (Fig. 1a). Annual precipitation is 550e600 mm. The soil is classified as an Earth-cumuli-Orthic Anthrosol, with a silt loam texture in the surface layer. Selected properties of the surface layer (0e20 cm) were: sand 31.8%, silt 51.7%, clay 16.5%, bulk density 1.33 g cm3, organic C 9.83 g kg1, total N 0.88 g kg1, inorganic e 1 N (NHþ 4 eN and NO3 eN) 7.02 mg kg , pH (soil:water; 1:2.5) 8.3, field capacity 24.4%. During 2011e2012 and 2012e2013 season, the mean air temperature was 8.2 and 8.0  C; and precipitation totaled 112.9 and 191.8 mm, respectively. Before the study, a winter wheat (Triticum aestivum L.) e maize (Zea mays L.) rotation had been used for two years without irrigation and fertilization. Winter wheat cultivar was T. aestivum L. cv. Xiaoyan22. Nitrogen fertilizer was urea (46.6% N), while P fertilizer was single super phosphate (16% P2O5). The film was 0.001 cm thick, transparent, and made of polyethylene. Maize stalk was cut into small pieces using a stalk cutter, with selected properties as: average length 3 cm; moisture 142.7%; organic C 405.5 g kg1; total N 10.5 g kg1. 2.2. Experimental design Twelve treatments were arranged in a split-plot design with three replicates (Fig. 1b). Six tillage systems were established as the main plots, i.e., conventional tillage (CT, as the control), stalk mulching (SM), film mulching (FM), ridge tillage (RT), ridge tillage with film mulch on the ridge (RTfm), and ridge tillage with film mulch on the ridge and stalk mulch in the furrow (RTfmsm). Subplots consisted of two nitrogen application rates, i.e., 0 (N0) and 180 kg N ha1 [N180, the recommended nitrogen dose for winter wheat (Guo et al., 2012)]. Each plot was 8  4 m2 with 1.5 m gaps between blocks. The field was plowed with a rotary cultivator to 20 cm depth before sowing. In FM, both film and planted rows were 30 cm wide; while in RT, RTfm, and RTfmsm, both ridges and planted rows were 30 cm wide, and ridges were 15 cm in height (Fig. 1c). In RTfm and RTfmsm, ridges were mulched with film. In SM, plots were mulched with stalk at 5000 kg ha1; while in RTfmsm, the planted rows were mulched with stalk at 2500 kg ha1 (Fig. 1c). Fertilizers were applied as a basal fertilization before sowing wheat. The nitrogen application rate was 0 or 180 kg N ha1, while phosphorus application rate was 120 kg P2O5 ha1 for all treatments. In CT and SM, fertilizers were broadcast and then incorporated into 0e15 cm soil layer; while in the other four tillage systems, fertilizers were applied by deep-band application into 15 cm soil layer (Fig. 1c). Wheat was sown in the planted rows 20 cm apart in CT and SM, while 30 cm apart in the other four tillage

61

systems (Fig. 1c). Wheat was sown to a 5 cm depth at 120 kg ha1 for all treatments. No irrigation and side-dress fertilizer was applied during wheat growing seasons.

2.3. Ammonia measurement In current study, ammonia emissions were measured using a vented chamber method (Ma et al., 2010; Mkhabela et al., 2008). It is a recognized technique for measurement of ammonia emissions, especially suitable for field experiments with multiple treatments due to its simple structure, easy operation, and high recovery of ammonia (Wang et al., 2004; Zhang et al., 2014). Each vented chamber consisted of one cylindrical chamber (20 cm in height with a 15 cm internal diameter, white, made of polyvinyl chloride), two absorbers (5 cm thick with a 15 cm diameter, white, made of polyurethane), four poles (15 cm long with a 0.5 cm diameter, made of bamboo), and one rainproof (30  30 cm2, 0.5 cm thick, white, made of polyvinyl chloride) (Fig. 2). The absorbers were impregnated with 20 mL of a 0.8 mol L1 phosphoric acid (H3PO4) þ 0.7 mol L1 glycerol solutions to trap ammonia. The bottom absorber was 5 cm above the soil surface, used for trapping ammonia volatilized from the soil, while the top absorber used for eliminating background air ammonia interference on the bottom absorber. Only the bottom absorbers were analyzed. Ammonia trapped with bottom absorbers were extracted with 50 mL 1 mol L1 potassium chloride (KCl) solution, and then the extracts were analyzed using a continuous flow analyzer. After extraction, the absorbers were rinsed several times (3) with 0.8 mol L1 H3PO4 and distilled water for reuse. After fertilization, vented chambers were inserted into the soil to a 5 cm depth in the middle of each subplot (Fig. 2). In both N180 and N0 treatments, one chamber was placed in each plot in CT and SM; while two chambers, 30 cm apart from each other, were placed in each plot in the other four tillage systems, with one chamber placed in the fertilized row, and the other in the planted row. Film mulches beneath the chambers were removed. Ammonia emissions were measured at 2e6 d intervals. A recovery test was conducted to assess the ammonia-collecting efficiency of the vented chambers following the method reported by Wang et al. (2004). Our findings showed the average recovery rate was 99.0 ± 1.6% (n ¼ 5). As film beneath the chamber was removed, some effects of film on ammonia emissions were eliminated. Therefore, the results from film mulched treatments, i.e., FM, RTfm, and RTfmsm, may have been overestimated.

Fig. 2. Vented chamber for trapping ammonia.

62

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

2.4. Calculation equations for ammonia losses Ammonia losses in CT and SM were given by:


A  10000 Ammonia flux kg N ha1 d1 ¼ i S  Di  0:99 
Mean ammonia flux kg N ha1 d1 , X n  n X Ai  10000 ¼ Di S  0:99 i¼1 i¼1 
Cumulative ammonia emission CAE; kg N ha1  n  X Ai  10000 ¼ S  0:99 i¼1

(1)

(2)

(3)

Ammonia losses in FM, RT, RTfm, and RTfmsm were given by:




1

Ammonia flux kg N ha

1

d



ðAFi þ APi Þ=2  10000 ¼ S  Di  0:99


Mean ammonia flux kg N ha1 d1 , X n  n X ðAFi þ APi Þ=2  10000 ¼ Di S  0:99 i¼1 i¼1 
Cumulative ammonia emission CAE; kg N ha1  n  X ðAFi þ APi Þ=2  10000 ¼ S  0:99 i¼1

LAI ¼ leaf area (cm2 plant1) * plant density (plants m2) * 104(9) where, 104 is used to convert cm2 to m2.

(4) 2.6. Statistical analysis

(5)

(6)

where, Ai is ammonia loss from CT and SM for the ith sampling (kg N); AFi and APi is ammonia loss from the fertilized row and planted row in FM, RT, RTfm, and RTfmsm for the ith sampling, respectively (kg N); i represents the ith sampling; S is the trapping area of the vented chamber (0.0177 m2); Di is trapping duration (d) for the ith sampling; n is the number of sampling events each season (n ¼ 25 in the 2011e2012 season, and n ¼ 27 in the 2012e2013 season); ammonia recovery in the vented chamber is 99.0%. Fertilizer ammonia loss and fertilizer ammonia loss rate were given by: Fertilizer ammonia loss (FAL, kg N ha1) ¼ CAEN180 e CAEN0

oven at 105  C (Margesin and Schinner, 2005). Organic carbon was measured by titration using ferrous iron (Fe2þ) solution after wet oxidation with potassium dichromate (K2Cr2O7) and sulfuric acid (H2SO4) (Pansu and Gautheyrou, 2003). Soil pH was monitored using a 1:2.5 (w/v) soil/water solution with a pH meter (Margesin and Schinner, 2005). To measure ammonium, 20 g of each soil sample was extracted with 200 mL 1 mol L1 potassium chloride (KCl) solution, and then the extracts were analyzed using a continuous flow analyzer (Han et al., 2014a). Soil surface temperature (at 0 cm depth) was measured using an infrared thermometer. Soil temperature (at 2, 5, 10, 15, 20, and 25 cm depths) was measured using right-angled thermometers. Leaf area (cm2 plant1, n ¼ 20) was measured using the ImageJ image analysis system (Martin et al., 2013). Plant density (plants m2, n ¼ 3) was measured by counting plants in a 1  1 m2 area in each plot. Leaf area index (LAI) was given by:

(7)

Fertilizer ammonia loss rate (FALR, %) ¼ (CAEN180 e CAEN0)  100%/ FN (8) where, CAEN180 and CAEN0 is cumulative ammonia emission (kg N ha1) in N180 and N0 treatment, respectively; FN is fertilizer nitrogen application rate (180 kg N ha1). 2.5. Measurement of soil factors and leaf area index Three soil samples (0e20 cm depth) were collected using a steel core sampler (2.5 cm i.d.) from each subplot, and then mixed thoroughly for measurement of soil factors such as moisture, organic carbon, pH, and ammonium (NHþ 4 ) at 2e6 d intervals, consistent with the sampling of ammonia. Both at the beginning and in the end of each season, three soil samples (0e200 cm depth, 20 cm each layer) were collected using a steel core sampler (4 cm i.d.) from each subplot, and then mixed thoroughly for measurement of soil moisture. Soil moisture was measured using a drying

Data were evaluated by analysis of variance (ANOVA). Means were compared using the least significant difference (LSD) test. Regression analysis (RA) and multiple regression analysis (MRA) (Moore et al., 2009) were performed to fit the relationship between ammonia flux and soil factors as: y ¼ a þ b$x

(10)

y ¼ a þ bT$xT þ bm$xm þ ba$xa þ bpH$xpH þ bOC$xOC

(11)

where, y is ammonia flux; a is the intercept; x is soil factor; b is the regression coefficient for soil factor; bT, bm, ba, bpH, and bOC is the regression coefficient for soil temperature (xT), moisture (xm), ammonium (xa), pH (xpH), and organic carbon (xOC), respectively. To remove the impacts of different units on the comparison of different variables, data were normalized before RA and MRA by: xn ¼ (xo  xmin)/(xmax  xmin)

(12)

where, xn is the normalized value; xo is the original value; xmin and xmax is the minimum and maximum value, respectively. The ANOVA, RA, and MRA were conducted using SAS 9.1 (SAS Institute Inc., NC, US). 3. Results and discussion 3.1. Ammonia losses 3.1.1. Ammonia losses affected by tillage practices Tillage had significant impacts on ammonia emissions (P < 0.01) (Table 1). In N180 treatments, FM had the lowest ammonia flux peak, followed by RT, RTfm, RTfmsm, and SM, while CT had the greatest peak, i.e., 0.129 and 0.139 kg N ha1 d1 in 2011e2012 and 2012e2013 season, respectively (P > 0.05) (Fig. 3). Meanwhile, in N180 treatment, ammonia fluxes in most conservation tillage systems were lower than that in CT especially during Oct.eDec. (P < 0.05) (Fig. 3). Consequently, in N180 treatments, conservation tillage systems resulted in lower ammonia emissions, with the lowest observed in FM (5.27 kg N ha1) (Table 1). Moreover, in N180 treatments, ammonia emissions in the planted rows significantly varied among tillage systems during Oct.eDec. (P < 0.05), with the lowest also observed in FM (2.33 kg N ha1) (Table 2). While in N0

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

treatments, there was no significant effect of conservation tillage on ammonia losses (Tables 1 and 2). Averaged across N180 and N0 treatments, mean ammonia fluxes in conservation tillage systems ranged from 0.0211 to 0.0246 kg N ha1 d1, with the lowest observed in FM (0.0211 kg N ha1 d1), which were 18.2% lower than that in CT (0.0258 kg N ha1 d1) (Fig. 3). Moreover, FM also resulted in the lowest fertilizer ammonia loss in the planted rows during Oct.eDec. (0.10 kg N ha1) and in the fertilized rows during Feb.eMay (0.23 kg N ha1) (Table 2). Consequently, conservation tillage systems reduced ammonia losses, with FM most effective, which on average had a 1.09 kg N ha1 ammonia reduction compared with CT (Table 1). Then, by adopting FM, ammonia emission is expected to be reduced by 2.71  107 kg N each winter wheat growing season since winter wheat production systems occupy 2.49  107 ha of cultivated land in the southern Loess Plateau. However, this expected ammonia abatement needs further verifications with multi-location tests.

3.1.2. Ammonia losses affected by nitrogen fertilization Nitrogen fertilization had significant impacts on ammonia emissions (P < 0.01) (Table 1). Among tillage systems, ammonia losses in N80 treatments ranged from 5.27 to 7.23 kg N ha1, greater than that in N0 treatments (4.52e5.28 kg N ha1) (Table 1). In general, ammonia emissions in N180 treatments averaged 6.26 kg N ha1, 26.5% greater than that in N0 treatments (4.95 kg N ha1) (P < 0.05) (Table 1). Moreover, in the fertilized rows, ammonia losses in N180 treatments were greater than that in N0 treatments (P < 0.05); while in the planted rows ammonia losses in N180 treatments were also greater (P > 0.05) (Table 2), indicating effective management strategies should be implemented to reduce ammonia emissions resulted from nitrogen fertilization.

Table 1 Effects of growing season, tillage, and nitrogen on ammonia losses (kg N ha1). Source

F value

Pr > F

Source

F value

Pr > F

Season Tillage Nitrogen

0.39 24.72 171.81

0.59 <0.01 <0.01

Season  tillage Season  nitrogen Tillage  nitrogen

1.20 0.02 7.37

0.35 0.90 <0.01

Season

Average NH3 loss* (n ¼ 36)

Nitrogen

Average NH3 loss* (n ¼ 36)

2011e2012 2012e2013 Mean LSD

5.40 ± 1.30a 5.82 ± 0.87a 5.61 2.90

N180 N0 Mean LSD

6.26 ± 1.07a 4.95 ± 0.73b 5.61 0.21

Tillage

Average NH3 loss* NH3 loss in N180 NH3 loss in N0 FAL (n ¼ 12) (n ¼ 6) (n ¼ 6)

CT SM FM RT RTfm RTfmsm Mean LSD

5.99 5.69 4.90 5.61 5.73 5.73 5.61 0.22

± ± ± ± ± ±

1.58a 1.23b 0.82c 0.85b 1.05b 0.92b

7.23 6.51 5.27 6.04 6.35 6.17 6.26 0.37

± ± ± ± ± ±

1.13a 1.05b 0.78d 0.81c 0.98bc 0.94bc

4.74 4.87 4.52 5.19 5.12 5.28 4.95 0.38

± ± ± ± ± ±

0.73cd 0.76bcd 0.73d 0.72ab 0.74abc 0.69a

2.49 1.64 0.75 0.85 1.23 0.89 1.31 e

FALR, % 1.38 0.91 0.42 0.47 0.68 0.49 0.73 e

Tillage treatments consisted of conventional tillage (CT, as the control), stalk mulching (SM), film mulching (FM), ridge tillage (RT), ridge tillage with film mulch on the ridge (RTfm), and ridge tillage with film mulch on the ridge and stalk mulch in the furrow (RTfmsm). Nitrogen treatments consisted of 0 (N0) and 180 kg N ha1 (N180). Values are mean ± SD. Means with the same letter in the same column (for the same factor) are not significantly different (P > 0.05). LSD represents the least significant difference (a ¼ 0.05). * The values are the mean averaged across all treatments related to this factor. Fertilizer ammonia loss (FAL, kg N ha1) ¼ NH3 loss in N180 e NH3 loss in N0. Fertilizer ammonia loss rate (FALR, %) ¼ FAL  100%/N application rate.

63

3.1.3. Ammonia losses in different stages Ammonia losses varied with measurement stages. Ammonia fluxes peaked only during Oct.eDec., i.e., ammonia fluxes began to increase within 2e4 days after fertilization (DAF), reached a peak within 10 DAF, and then dropped rapidly and leveled off after 30 DAF (Fig. 3). This finding was similar to that reported by Gong et al. (2013b) and Shangguan et al. (2012). Consequently, during Oct.eDec., ammonia losses in N0 and N180 treatments averaged 2.68 and 3.30 kg N ha1, respectively, 14.0% and 23.6% greater than that during Feb.eMay, respectively (P < 0.05) (Table 2). Moreover, fertilizer ammonia loss during Oct.eDec. averaged 0.62 kg N ha1, 93.8% greater than that during Feb.eMay (0.32 kg N ha1) (P < 0.05) (Table 2). All those indicated more ammonia losses occurred during Oct.eDec. One explanation is a great amount of ammonium formed from urea hydrolysis within 10 DAF, leading to an ammonia emission peak in N180 treatment (Shangguan et al., 2012). Furthermore, soil tillage loosened the top soil layer, and then promoted organic nitrogen mineralization which produced more ammonium; meanwhile, the resistance of the top soil layer to ammonia diffusion was reduced. All those together led to an ammonia emission peak after sowing even in N0 treatments (Fig. 3). 3.1.4. Ammonia losses in different placements Ammonia losses varied among measurement placements. For instance, in N180 treatments during Oct.eDec., ammonia emissions in the fertilized rows averaged 3.70 kg N ha1, 28.0% greater than that in the planted rows (2.89 kg N ha1) (P < 0.05); while the mean fertilizer ammonia loss in the fertilized rows was 0.97 kg N ha1, 288.0% greater than that in the planted rows (0.25 kg N ha1) (P < 0.05) (Table 2). All those indicated more ammonia emissions occurred in the fertilized rows. 3.1.5. Ammonia losses in current and previous studies Current study found 2.49 kg N ha1 loss (1.38% of applied N) in conventional tillage in the southern Loess Plateau of China (Table 1), lower than some previous findings. For example, Engel et al. (2011) reported 20.5 kg N ha1 loss (20.5% of applied N) in the semiarid northern Great Plains of the United States; Datta et al. (2012) found 12 kg N ha1 loss (10% of applied N) in the IndoGangetic alluvial of India; Turner et al. (2010) observed 7.6 kg N ha1 loss (9.5% of applied N) in western Victoria Australia. However, there are also some findings similar to current findings. For instance, Shangguan et al. (2012) reported 3.39 kg N ha1 loss (1.9% of applied N) in the southern Loess Plateau of China. Ma et al. (2012b) found 2.21 kg N ha1 loss (1.5% of applied N) in Hebei of China. Phosphorus fertilization (pH 2.3) is partially responsible for low ammonia losses in current study as it stabilized soil ammonium and thus restricted ammonia volatilization (Akhtar and Naeem, 2012). Besides, conservation tillage practices further reduced ammonia losses in current study. 3.2. Relationships between ammonia fluxes and soil factors Effects of soil factors on ammonia fluxes changed with stages and nitrogen treatments (Table 3). During Oct.eDec., ammonia fluxes in N0 treatments were mainly related to organic carbon, pH, temperature, and ammonium; while in N180 treatments mainly related to ammonium and pH. During Feb.eMay, ammonia fluxes in N0 treatments were mainly affected by temperature, moisture, and ammonium; while in N180 treatments mainly affected by temperature, moisture, ammonium, and organic carbon. In general, ammonia fluxes were mainly affected by soil ammonium, moisture, and temperature (Table 3).

64

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

Fig. 3. Variations of ammonia fluxes during 2011e2013 winter wheat growing seasons. Error bars indicate standard deviation (n ¼ 3).

Table 2 Ammonia emissions and fertilizer ammonia loss (FAL, kg N ha1) in different placements, tillage systems, nitrogen treatments, and stages. Stage

Oct.eDec.

Placement

Tillage

N0

Fertilized row

FM RT RTfm RTfmsm LSD Mean (n ¼ 24) LSD for Mean FM RT RTfm RTfmsm LSD Mean (n ¼ 24) LSD for Mean

2.45 2.84 2.73 2.89 0.72 2.73 0.30 2.23 2.88 2.67 2.76 0.62 2.64 0.57 2.68 0.42

Planted row

Grand mean (n ¼ 48) LSD for grand mean

Feb.eMay N180

0.48a 0.62a 0.68a 0.58a

0.84a 0.93a 0.85a 0.91a

3.50 3.69 3.74 3.86 1.07 3.70

± ± ± ±

± ± ± ±

± 0.54A

2.33 3.11 2.96 3.17 0.64 2.89

± 0.56B

3.30 ± 0.83A

± ± ± ±

± 0.58B ± ± ± ±

0.53a 0.54a 0.55a 0.43a

± 0.84A 0.50b 0.51a 0.47ab 0.62a

± 0.60A

FAL

N0

1.05 0.85 1.01 0.97 e 0.97 e 0.10 0.23 0.29 0.41 e 0.25 e 0.62 e

2.20 2.28 2.41 2.42 0.46 2.33 0.35 2.17 2.37 2.43 2.50 0.57 2.37 0.36 2.35 0.21

N180 ± ± ± ±

0.37a 0.37a 0.40a 0.38a

FAL 0.43b 0.41ab 0.58a 0.45ab

2.43 2.67 3.07 2.71 0.57 2.72

± ± ± ±

± ± ± ±

± 0.46A

2.29 2.62 2.92 2.61 0.59 2.61

± 0.41B

2.67 ± 0.50A

± 0.37B ± ± ± ±

0.35a 0.56a 0.44a 0.50a

± 0.50A 0.40a 0.57a 0.51a 0.47a

± 0.51A

0.23 0.39 0.66 0.29 e 0.39 e 0.12 0.25 0.49 0.11 e 0.24 e 0.32 e

Tillage treatments consisted of film mulching (FM), ridge tillage (RT), ridge tillage with film mulch on the ridge (RTfm), and ridge tillage with film mulch on the ridge and stalk mulch in the furrow (RTfmsm). Nitrogen treatments consisted of 0 (N0) and 180 kg N ha1 (N180). Values are mean ± SD (n ¼ 6). Mean is the value averaged across tillage systems in the same fertilizer placement. Grand mean is the value averaged across tillage systems and fertilizer placements. Means with the same lowercase letters in the same column and fertilizer placement are not significantly different (P > 0.05); means with the same capital letters in the same row and stage are not significantly different (P > 0.05). Grey background color indicates significant differences were observed in the means between nitrogen treatments or among tillage systems (P < 0.05). LSD represents the least significant difference (a ¼ 0.05). FAL ¼ NH3 loss in N180 e NH3 loss in N0.

3.2.1. Ammonia fluxes versus soil ammonium Ammonia fluxes varied with soil ammonium concentration. In most cases, soil ammonium had significantly positive effects on ammonia fluxes (Fig. 4). During Oct.eDec. and Feb.eMay, soil ammonium explained 15.6e33.2% and 0.1e22.6% of the variation in ammonia fluxes, respectively. Zhang et al. (2014) and Han et al. (2014b) also found ammonia fluxes significantly increased as nitrogen input increased. During Oct.eDec., responses of ammonia

fluxes to ammonium in conservation tillage systems were weaker than in conventional tillage (Fig. 4), partially explained why ammonia losses were lower in conservation tillage systems (Table 1). 3.2.2. Ammonia fluxes versus soil moisture Ammonia fluxes changed with soil moisture. Soil moisture had significantly positive effects on ammonia fluxes (Fig. 4). During

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

65

Table 3 Regression coefficients from multiple regression analysis between ammonia flux and soil factors in different stages and nitrogen treatments. Oct.eDec.

a bT bm ba bpH bOC

Oct.eDec. & Feb.eMay

Feb.eMay

N0 (n ¼ 138)

N180 (n ¼ 138)

N0 (n ¼ 174)

N180 (n ¼ 174)

N0 & N180 (n ¼ 624)

0.13 0.31z ns 0.20z 0.49z 0.45z

0.002 ns 0.08 0.70z 0.17y ns

0.40z 0.41z 0.24z 0.39z ns ns

0.43z 0.25z 0.21z 0.34z ns 0.13y

0.04 0.08y 0.34z 0.36z ns ns

Nitrogen treatments consist of 0 (N0) and 180 kg N ha1 (N180). ns P > 0.15; yP < 0.05; zP < 0.01.

Oct.eDec. and Feb.eMay, soil moisture explained 6.4e17.5% and 8.2e21.9% of the variation in ammonia fluxes, respectively. Bosch-

Serra et al. (2014) found significant ammonia reductions when moisture of 0e30 cm soil layer was less than 56% of the field

Fig. 4. Relationship between ammonia flux and soil ammonium, moisture, and temperature in different tillage systems. Values have been normalized before the regression. In tillage system with a regression line above others, the response of ammonia flux to the soil factor is stronger. *P < 0.05; **P < 0.01.

66

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

capacity. Proctor et al. (2010) observed ammonia fluxes usually increase as soil moisture increases, but also with exceptions. For example, though heavy rainfall increase soil moisture, it does not increase ammonia fluxes as it promoted nitrogen fertilizer leaching into deeper soil layers (Proctor et al., 2010; Sanz-Cobena et al., 2011; Han et al., 2014b). During Oct.eDec., responses of ammonia fluxes to moisture were weaker in conservation tillage systems than in conventional tillage (Fig. 4), partially explained why ammonia losses were lower in conservation tillage systems (Table 1).

3.2.3. Ammonia fluxes versus soil temperature Ammonia fluxes varied with soil temperature. Soil temperature had positive effects on ammonia fluxes during Oct.eDec., but had negative effects during Feb.eMay (Fig. 4). One explanation is during Oct.eDec., initial moisture of the 0e2 m soil profile was relatively high (17e22%), while air temperature (averaged 4.1  C), soil temperature (averaged 13.4  C), and wheat leaf area index (averaged 1.09) was relatively low, leading to little soil moisture loss via evaporation and transpiration (Fig. 5). But relatively high soil temperature promoted the formation of ammonium from urea and organic nitrogen mineralization and then increased ammonia fluxes (Fan et al., 2011; Shangguan et al., 2012). Therefore, soil temperature had a positive effect on ammonia flux during Oct.eDec. (Fig. 4). Moreover, during Oct.eDec., responses of ammonia fluxes to temperature were weaker in conservation tillage systems than in conventional tillage (Fig. 4), partially explained why ammonia losses were lower in conservation tillage systems (Table 1). While during Feb.eMay, air temperature, soil temperature, and leaf area index gradually increased, and soil moisture decreased to 12e16% via evaporation and transpiration (Fig. 5), which limited the formation of ammonium from organic nitrogen mineralization (Shangguan et al., 2012), and then decreased ammonia fluxes. Therefore, ammonia flux was negatively correlated with soil temperature during Feb.eMay (Fig. 4). During Oct.eDec. and Feb.eMay, soil temperature explained

18.3e34.3% and 10.1e28.6% of the variation in ammonia flux, respectively. Fan et al. (2011) and Shangguan et al. (2012) also found ammonia fluxes were strongly related to soil temperature. 3.3. Mechanisms of ammonia loss variations among tillage systems 3.3.1. Effects of deep-band application of nitrogen fertilizer Soil ammonium had a positive effect on ammonia flux (Fig. 4). Deep-band application of nitrogen fertilizers did not directly increased soil ammonium concentration in the top soil layer (0e15 cm), but restricted gas exchange between the soil and the atmosphere. Therefore, ammonia emissions decreased in the four conservation tillage systems which applied nitrogen by deep-band application, i.e., FM and ridge tillage systems (Fig. 1c, Table 1), in which, ridge tillage systems had larger ammonia losses, partially due to their greater soil surface in the fertilized rows (ridges) for ammonia emissions (Fig. 1c). 3.3.2. Effects of stalk mulching Soil temperature had a positive effect on ammonia flux (Fig. 4). In stalk mulching (SM) treatment, fertilizer placement method (Fig. 1c) and ammonia emission peaks (Fig. 3) was similar to that of conventional tillage. Therefore, SM was not expected to reduce ammonia emissions. However, current study showed the cumulative ammonia losses in SM was lower than that in conventional tillage (Table 1). One explanation is the mechanisms of ammonia abatement in SM were different from the treatments applied nitrogen by deep-band application such as ridge tillage (Fig. 1c). Stalk mulches usually acted as heat-insulating materials on the soil surface, thus alleviated solar radiations and then decreased soil temperature (Fig. 6), leading to reductions in ammonia emissions (Fan et al., 2011). 3.3.3. Effects of film mulching Film mulches could increase soil temperature and moisture

Fig. 5. Air temperature, soil temperature (0e20 cm profile), soil moisture (0e2 m profile), and wheat leaf area index (LAI) during wheat growing season. Values of soil temperature, moisture and LAI were the means averaged among tillage systems.

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

67

Fig. 6. Average soil temperature and moisture (0e20 cm profile, n ¼ 104) in different placements. Means with the same letter are not significantly different (P > 0.05).

(Fig. 6), and therefore increased ammonia emission potentials. However, as film mulches restricted the gas flow between the soil and the atmosphere, they could also decrease ammonia emissions (Shangguan et al., 2012). Current study has partially overestimated the ammonia losses in film mulched treatments, i.e., FM, RTfm, and RTfmsm. Therefore, these conservation tillage systems may be more significant in reducing ammonia losses, further verifying our conclusion that conservation tillage practices such as film mulching could alleviate ammonia emissions. 3.3.4. Effects of higher moisture in conservation tillage systems Soil moisture had a positive effect on ammonia flux (Fig. 4). Most conservation tillage systems increased soil moisture (Fig. 6). Thus, conservation tillage systems were expected to increase ammonia losses. In fact, in N0 treatments, ammonia losses in most conservation tillage systems were higher than that in conventional tillage, though the differences were not significant (P > 0.05) (Table 1). These findings also implied that conservation tillage systems reduce ammonia losses from fertilizer nitrogen rather than from soil nitrogen. 4. Conclusions Nitrogen application significantly increased ammonia emission. Ammonia fluxes were mainly related to soil ammonium, moisture, and temperature in this region, and responses of ammonia fluxes to these soil factors were weaker in conservation tillage systems than in conventional tillage. Tillage systems had significant effects on ammonia emissions. Conservation tillage practices reduced ammonia emissions from fertilizer nitrogen, with film mulching (FM) most effective. Deep-band application of nitrogen fertilizer, stalk mulches, and film mulches were responsible for reductions in ammonia emissions from nitrogen fertilization in conservation tillage systems, thus they were recommended to reduce ammonia emissions from winter wheat production regions in the southern Loess Plateau. Acknowledgments This study was supported by Agricultural Scientific Research

Talent and Team Program (2013) of the Ministry of Agriculture, China. We thank Mr. Anmin Xu and Yongyi Wang for laboratory assistance, and Prof. Mancai Guo, Yingzhou Liu, and Yuncheng Wang for their help with statistical analysis. References Akhtar, M., Naeem, A., 2012. Reduction in ammonia loss by applying urea in combination with phosphate sources. Commun. Soil Sci. Plan. 43, 2043e2049. Behera, S.N., Sharma, M., Aneja, V.P., Balasubramanian, R., 2013. Ammonia in the atmosphere: a review on emission sources, atmospheric chemistry and deposition on terrestrial bodies. Environ. Sci. Pollut. Res. 20, 8092e8131. Bosch-Serra, A.D., Yague, M.R., Teira-Esmatges, M.R., 2014. Ammonia emissions from different fertilizing strategies in Mediterranean rainfed winter cereals. Atmos. Environ. 84, 204e212. Datta, A., Sharma, S.K., Harit, R.C., Kumar, V., Mandal, T.K., Pathak, H., 2012. Ammonia emission from subtropical crop land area in India. Asia Pac. J. Atmos. Sci. 48, 275e281. Engel, R., Jones, C., Wallander, R., 2011. Ammonia volatilization from urea and mitigation by NBPT following surface application to cold soils. Soil Sci. Soc. Am. J. 75, 2348e2357. Fan, X.H., Li, Y.C., Alva, A.K., 2011. Effects of temperature and soil type on ammonia volatilization from slow-release nitrogen fertilizers. Commun. Soil Sci. Plan. 42, 1111e1122. Ferm, M., 1998. Atmospheric ammonia and ammonium transport in Europe and critical loads: a review. Nutr. Cycl. Agroecosystems 51, 5e17. Gong, L.W., Lewicki, R., Griffin, R.J., Tittel, F.K., Lonsdale, C.R., Stevens, R.G., Pierce, J.R., Malloy, Q.G.J., Travis, S.A., Bobmanuel, L.M., Lefer, B.L., Flynn, J.H., 2013a. Role of atmospheric ammonia in particulate matter formation in Houston during summertime. Atmos. Environ. 77, 893e900. Gong, W.W., Zhang, Y.S., Huang, X.F., Luan, S.J., 2013b. High-resolution measurement of ammonia emissions from fertilization of vegetable and rice crops in the Pearl River Delta Region, China. Atmos. Environ. 65, 1e10. Guo, S.L., Zhu, H.H., Dang, T.H., Wu, J.S., Liu, W.Z., Hao, M.D., Li, Y., Syers, J.K., 2012. Winter wheat grain yield associated with precipitation distribution under longterm nitrogen fertilization in the semiarid Loess Plateau in China. Geoderma 189, 442e450. Han, K., Yang, Y., Zhou, C.J., Shangguan, Y.X., Zhang, L., Li, N., Wang, L.Q., 2014a. Management of furrow irrigation and nitrogen application on summer Maize. Agron. J. 106, 1402e1410. Han, K., Zhou, C.J., Wang, L.Q., 2014b. Reducing ammonia volatilization from Maize fields with separation of nitrogen fertilizer and water in an alternating furrow irrigation system. J. Integr. Agric. 13, 1099e1112. Liu, X., Zhang, Y., Han, W., Tang, A., Shen, J., Cui, Z., Vitousek, P., Erisman, J.W., Goulding, K., Christie, P., Fangmeier, A., Zhang, F., 2013. Enhanced nitrogen deposition over China. Nature 494, 459e462. Ma, B.L., Wu, T.Y., Tremblay, N., Deen, W., McLaughlin, N.B., Morrison, M.J., Stewart, G., 2010. On-farm assessment of the amount and timing of nitrogen fertilizer on ammonia volatilization. Agron. J. 102, 134e144. Ma, J.Z., Xu, X.B., Zhao, C.S., Yan, P., 2012a. A review of atmospheric chemistry research in China: photochemical smog, haze pollution, and gas-aerosol

68

Y. Yang et al. / Atmospheric Environment 104 (2015) 59e68

interactions. Adv. Atmos. Sci. 29, 1006e1026. Ma, Y., Ji, Y., Li, X., Zhang, L., Ju, X., Zhang, L., 2012b. Effects of N Fertilization rates on the NH3 volatilization and N2O emissions form the wheat-maize rotation system in North China Plain. Ecol. Environ. Sci. 21, 225e230 (in Chinese with English abstract). Margesin, R., Schinner, F., 2005. Manual of Soil Analysis: Monitoring and Assessing Soil Bioremediation. Springer, Heidelberg, pp. 47e71. Martin, T.N., Marchese, J.A., de Sousa, A.K.F., Curti, G.L., Fogolari, H., Cunha, V.D., 2013. Using the ImageJ software to estimate leaf area in bean crop. Interciencia 38, 843e848. Martinez-Lagos, J., Salazar, E., Alfaro, M., Misselbrook, T., 2013. Ammonia volatilization following dairy slurry application to a permanent grassland on a volcanic soil. Atmos. Environ. 80, 226e231. Mkhabela, M.S., Madani, A., Gordon, R., Burton, D., Cudmore, D., Elrni, A., Hart, W., 2008. Gaseous and leaching nitrogen losses from no-tillage and conventional tillage systems following surface application of cattle manure. Soil Tillage Res. 98, 187e199. Moore, D.S., McCabe, G.P., Craig, B.A., 2009. Introduction to the Practice of Statistics. W. H. Freeman and Company, New York, pp. 607e636. Pansu, M., Gautheyrou, J., 2003. Handbook of Soil Analysis: Mineralogical, Organic and Inorganic Methods. Springer, Heidelberg, pp. 340e341. Proctor, C., Koenig, R., Johnston, W., 2010. Potential for ammonia volatilization from urea in Dryland Kentucky bluegrass seed production systems. Commun. Soil Sci. Plan. 41, 320e331.

Roelcke, M., Han, Y., Li, S.X., Richter, J., 1996. Laboratory measurements and simulations of ammonia volatilization from urea applied to calcareous Chinese loess soils. Plant Soil 181, 123e129. Sanz-Cobena, A., Misselbrook, T., Camp, V., Vallejo, A., 2011. Effect of water addition and the urease inhibitor NBPT on the abatement of ammonia emission from surface applied urea. Atmos. Environ. 45, 1517e1524. Shangguan, Y.X., Shi, R.P., Li, N., Han, K., Li, H.K., Wang, L.Q., 2012. Factors influencing ammonia volatilization in a winter wheat field with plastic film mulched ridges and unmulched furrows. Environ. Sci. 33, 1987e1993 (in Chinese with English abstract). Turner, D.A., Edis, R.B., Chen, D., Freney, J.R., Denmead, O.T., Christie, R., 2010. Determination and mitigation of ammonia loss from urea applied to winter wheat with N-(n-butyl) thiophosphorictriamide. Agric. Ecosyst. Environ. 137, 261e266. Wang, Z.H., Liu, X.J., Ju, X.T., Zhang, F.S., Malhi, S.S., 2004. Ammonia volatilization loss from surface-broadcast urea: comparison of vented- and closed-chamber methods and loss in winter wheat-summer maize rotation in North China Plain. Commun. Soil Sci. Plan. 35, 2917e2939. Webb, J., Thorman, R.E., Fernanda-Aller, M., Jackson, D.R., 2014. Emission factors for ammonia and nitrous oxide emissions following immediate manure incorporation on two contrasting soil types. Atmos. Environ. 82, 280e287. Zhang, Y.H., Han, X., He, N.P., Long, M., Huang, J.H., Zhang, G.M., Wang, Q.B., Han, X.G., 2014. Increase in ammonia volatilization from soil in response to N deposition in inner Mongolia grasslands. Atmos. Environ. 84, 156e162.