Geoderma 433 (2023) 116460

Contents lists available at ScienceDirect

Geoderma
journal homepage: www.elsevier.com/locate/geoderma

Ammonia volatilization, greenhouse gas emissions and microbiological

mechanisms following the application of nitrogen fertilizers in a
saline-alkali paddy ecosystem
Xinyi Wang a, b, c, Hui Zhu a, c, *, Baixing Yan a, c, Lei Chen d, Brian Shutes e, Mingming Wang a, f,
Jiao Lyu a, c, Fuman Zhang a, c
a
Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun 130102, China
b
College of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100049, China
c
Jilin Provincial Engineering Center of CWs Design in Cold Region & Beautiful Country Construction, Changchun 130102, China
d
State Key Laboratory of Water Environment Simulation, School of Environment, Beijing Normal University, Beijing, China
e
Department of Natural Sciences, Middlesex University, Hendon, London NW4 4BT, UK
f
Da’an Sodic Land Experiment Station, Da’an, Jilin 131300, China

A R T I C L E I N F O A B S T R A C T

Handling Editor: Naoise Nunan The application of nitrogen (N) fertilizer can promote rice yield, but is accompanied by considerable quantities of
ammonia (NH3) volatilization and greenhouse gas (GHG) emissions. The effect of N-fertilization to these gas
Keywords: emissions in saline-alkali paddy ecosystems is unclear. A 137-day mesocosm-scale experiment was carried out to
Saline-alkali paddy field clarify the effect of different N-fertilizer application strategies on NH3 volatilization and GHG emissions in saline-
Nitrogen fertilizer
alkali paddy ecosystems and to determine the microbiological mechanisms. Five N-fertilizer treatments con­
Greenhouse gas
sisting of control without N-fertilizer (CK), urea (U), urea with urease-nitrification inhibitors (UI), organ­
Ammonia volatilization
Functional N-cycling genes ic–inorganic compound fertilizer (OCF) and carbon-based slow-release fertilizer (CSF) were established. During
the entire rice-growing season, the cumulative NH3 emissions of UI, OCF and CSF treatments were 22.60–25.55
% significantly (p < 0.05) higher than U, respectively. The cumulative methane emissions in all N-fertilizer
treatments were decreased by 9.23–28.24 % compared with CK. The cumulative carbon dioxide emissions were
U > CSF > OCF > CK > UI, and the cumulative nitrous oxide (N2O) emissions were U > CSF > UI > CK > OCF.
The global warming potential of UI was 5.39–32.35 % lower than all the other treatments. Compared with the
other N-fertilizer treatments, the relative abundance of N-related functional bacteria (e.g., Sphingomonas and
Alphaproteobacteria) in OCF treatment was increased, and the (nirS + nirK)/nosZ ratio was reduced, thereby
decreasing the N2O emission. Overall, U is a better choice for controlling NH3 volatilization in saline-alkali paddy
ecosystems, whereas UI and OCF are more beneficial for reducing GHG emissions.

1. Introduction but also cause various environmental problems, e.g., soil nutrient
imbalance, soil acidification and salinization, surface and groundwater
Rice is one of the most important food crops in the world, and pollution, and especially ammonia (NH3), nitrogen oxides (NOx) and
increasing rice production is the primary task of ensuring world food other gases emission into the atmosphere (Guo et al. 2021, Sun et al.
security (Riya et al. 2012, Sun et al. 2018). In order to increase the rice 2021, Zhang et al. 2019, 2021).
yield, the amount of nitrogen (N) fertilizer applied per unit area of The NH3 volatilization from paddy fields is affected by the N-fertil­
paddy fields in the world has been increasing year by year, exceeding the izer types, resulting in different NH3 volatilization rates (Guo et al.
level required by crops and causing serious N losses (Corrochano-Mon­ 2021). The NH3 volatilized from farmland fertilization is one of the main
salve et al. 2021, Zhong et al. 2021). Inappropriate and excessive sources of atmospheric NH3 or ammonium (NH+ 4 ) (Zhang et al. 2019). As
application of N-fertilizer not only lead to a decrease in N utilization rate a main alkaline gas in the atmosphere, NH3 can react with sulfur dioxide

* Corresponding author at: Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences,
Changchun 130102, China.
E-mail address: [email protected] (H. Zhu).

https://doi.org/10.1016/j.geoderma.2023.116460
Received 6 October 2022; Received in revised form 26 March 2023; Accepted 29 March 2023
Available online 5 April 2023
0016-7061/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
X. Wang et al. Geoderma 433 (2023) 116460

and NOx to generate secondary particles, e.g., ammonium nitrate and 2. Material and methods
ammonium sulfate, which are important components of PM2.5 (atmo­
spheric particulate matter 2.5 µm or less in diameter) (Pan et al. 2018). 2.1. Experimental site
NH3 is therefore considered to be one of the key prerequisite substances
for atmospheric aerosol formation (Zhang et al. 2019). The NH3 vola­ A total of fifteen paddy field mesocosms were established under the
tilization process is rapid, which generally occurs within one week after mobile canopy awning at the Northeast Institute of Geography and
the application of N-fertilizers, and is directly and/or indirectly affected Agroecology, Chinese Academy of Sciences located in Changchun, China
by various factors (Sun et al. 2018, Wang et al. 2021a). The physical and (44◦ 0′ 24′′ N, 125◦ 24′ 53′′ E). The daily temperature and relative humidi­
chemical properties of soils (e.g., cation exchange capacity, pH and ty, during the 2021 rice-growing season (May 30th – October 13th) were
mechanical composition) affect the adsorption of NH+ 4 by soil colloids shown in Fig. S1 of Supplementary Materials. During the entire rice-
and the transformation from NH+ 4 to NH3 (Shan et al. 2015). The type growing season, the maximum and minimum of daily air temperatures
and dosage of N-fertilizers will also affect the NH+ 4 concentration of soil were 33.0 ℃ and 2.0 ℃, respectively, and the average daily air tem­
(Sun et al. 2018, Zhan et al. 2021). Meanwhile, greenhouse gas (GHG, i. perature was 20.4 ± 6.6 ℃. The average relative humidity was 83.51 ±
e., methane (CH4), carbon dioxide (CO2) and nitrous oxide (N2O)) 10.08 % and the annual precipitation was 860.3 mm.
emissions result from the comprehensive effects of N and carbon (C) According to the previous investigation (Yu et al. 2021, Yu et al.
coupling in paddy fields, and are considered to be a major contributor to 2023), the saline-alkali soil (0 – 20 cm topsoil) used for this study, which
global warming (Riya et al. 2012, Wang et al. 2021a). Application of N- is further classified as alkaline sodic soil, was collected at 9 randomly
fertilizer can promote rice growth, increase the input of exogenous C (e. selected sites from paddy fields in the hinterland of Songnen Plain in
g., root exudates, litter and stubble), and provide substrates for metha­ Da’an City, Western Jilin Province, China (45◦ 34′ 18–33′′ N,
nogens, thereby promoting the CH4 production (Yang et al. 2010). The 123◦ 54′ 25–42′′ E). Songnen Plain is one of the three major regions with
global warming potential (GWP) of CH4 is 34 times that of CO2 (Cao saline-alkali soils in the world, which has a large distribution area, low
et al. 2021). Meanwhile, the GWP of N2O, which is one of the main land productivity and a harsh ecological environment (Zhao et al.
causes of the greenhouse effect, is 298 times that of CO2 (Cao et al. 2020b). Prior to experimental operations, the soil from each site was
2021). A large amount of N2O is emitted from paddy fields in the N cycle pooled into a single sample, which was mixed evenly by using a concrete
process, which is mediated by ammonia-oxidizing bacteria and deni­ truck. The physicochemical properties of saline-alkali paddy soil were as
trifying bacteria in paddy soil (Braker and Conrad 2011). Ammonia follows: electrical conductivity (EC, ratio of soil to water was 1:10, w/v)
volatilization is increased with rising temperatures and can therefore be of 298.79 ± 128.58 μS cm− 1, pH (ratio of soil to water was 1:10, w/v) of
indirectly affected by warming caused by GHG emissions (Wang et al. 8.86 ± 0.41, soil organic matter (SOM) of 18.25 ± 2.10 g kg− 1, total N of
2021a). Therefore, it is necessary to implement a comprehensive 0.62 ± 0.20 g kg− 1, total phosphorus of 0.17 ± 0.04 g kg− 1, alkali-
assessment of NH3 volatilization and GHG emissions in paddy fields to hydrolyzable N of 42.88 ± 23.40 mg kg− 1, available phosphorus of
optimize fertilization management. 13.52 ± 6.81 mg kg− 1, and available potassium of 124.13 ± 29.14 mg
Around 9.56 × 108 ha of the world’s land is affected by salinization kg− 1.
(Wong et al. 2010). Due to poor soil quality, low yield and wide distri­
bution, saline-alkali land has become a worldwide problem of resource 2.2. Experimental design and agronomic management
and ecology, affecting about 50 % of farmland in the world (Hassani
et al. 2020, Qadir et al. 2007). Planting rice is one of the effective Each paddy field mesocosm was constructed using polyethylene
strategies to improve and utilize saline-alkali land (Chi et al. 2012). material (a cuboid of 640 mm length × 490 mm width × 360 mm
However, salinization deteriorates physical and chemical soil properties height). This experiment strictly followed the fertilizer application and
and reduces nutrient availability of N-fertilizer, thereby contributing to procedures management of local farmers in Western Jilin Province,
the N loss via different pathways (Chi et al. 2012, Shan et al. 2015). As China. Fertilizers were applied at the basal (BF), tillering (TF) and
one of the main pathways of N loss, the NH3 volatilization in non-saline- panicle initiation (PIF) growth stages. The percentages of N fertilizer
alkali paddy soil has been well studied (Wang et al. 2021a, Zhong et al. applied at BF, TF and PIF stages were 50 %, 30 % and 20 %, respectively.
2021), but environmental problems from the NH3 volatilization in full- Five treatments of N fertilization at BF stage were designed with three
scale saline-alkali paddy ecosystems are rarely investigated. Microor­ replicates for each treatment as follows: (1) control without N-fertilizer
ganisms are the main driving force of the C and N cycles in soil. High application (only 27.40 kg N ha− 1 from phosphate (P) fertilizer, i.e.,
salinity and alkalinity can affect the transformation process and ammonium phosphate, referred to CK), (2) urea (46 % N, referred to U),
migration behavior of C and N by inhibiting most microbial activity, (3) urea with 1 % N-(N-butyl)thiophosphoric triamide (NBPT) and 1 %
thus affecting the production and emission of NH3 and GHG (Jeppesen 3,4-dimethylpyrazole phosphate (DMPP) (46 % N, referred to UI), (4)
et al. 2020, Wang et al. 2020, Zhao et al. 2020a). Nowadays, saline- organic–inorganic compound fertilizer (N:P2O5:K2O = 12:0:3, referred
alkali land area is expanding in the world, and soil salinization is to OCF) and (5) C-based slow-release fertilizer (N:P2O5:K2O = 24:8:10,
increasing (Hassani et al. 2020). However, studies on NH3 volatilization referred to CSF). The total N-fertilizers were applied at the same amount
and/or GHG emissions from paddy fields are rarely carried out in saline- of 200 kg N ha− 1 in all treatments (excluding CK) during the entire rice-
alkali areas, and the relevant comprehensive assessment is not con­ growing season. All fertilizers were purchased from local regular fer­
ducted. Therefore, the study of both NH3 volatilization and GHG emis­ tilizer suppliers, of which the organic–inorganic compound fertilizer
sions in full-scale saline-alkali paddy fields is urgently needed, and the contained an additional 20 % organic matter and the C-based slow-
internal microbiological mechanisms need to be further discussed. release fertilizer contained 10 % biochar. Ammonium phosphate (N:
In order to fulfill the above knowledge gaps, four typical N-fertilizers P2O5:K2O = 18:46:0) and potassium sulfate (50 % K2O) were used as the
(i.e., urea, urea with urease-nitrification inhibitors, organic–inorganic P and potash (K) fertilizers, respectively. For a better yield increase, both
compound fertilizer and C-based slow-release fertilizer) were selected P (70 kg P2O5 ha− 1) and K (90 kg K2O ha− 1) fertilizers were applied
and applied through field surveys. A 137-day experiment which covered together with the initial dose of N-fertilizer at the BF stage before
an entire rice-growing season was conducted. The research objectives of transplanting rice. The BF was spread and then mixed completely with
this study are: (1) to investigate the variation of NH3 volatilization and the soil on May 30th, 2021. At the TF and PIF stages, 60 kg N ha− 1 and
GHG emissions under the application of various N-fertilizers in saline- 40 kg N ha− 1 of urea were used by sprinkling fertilizer over the fields,
alkali paddy ecosystems; and (2) to elucidate the microbiological respectively.
mechanisms and molecular drivers of gas emissions. All mesocosms were initially flooded (with an average water depth of
5 cm) on May 30th, and rice seedlings were transplanted on May 31st,

2
X. Wang et al. Geoderma 433 (2023) 116460

2021. The rice variety is Dongdao 4 (Oryza sativa L.), which is a proven Each chamber was a cuboid made of transparent polyethylene material.
high-yielding rice genotype in saline-alkali paddy fields in Northeast The air was pumped out for 4 h using an air-sampling pump (KVP04-
China (Li and Yang 2020). Eight holes of rice seedlings were trans­ 1.1–12, Kamoer) and allowed to pass through NH3 absorbent (500 mL of
planted in each mesocosm. Thereafter, all mesocosms were frequently 0.01 M H2SO4) contained inside a high-density polyethylene flask. Air-
flooded by intermittent irrigation before crop harvesting. All the treat­ sampling pumps were used to adjust the air velocity of NH3 collection
ments received the same agricultural managements (e.g., irrigation), in each chamber to ensure the accuracy of uniform collection of NH3
and there was no precipitation during the entire experimental period. emissions in all treatments. A cleaning bottle was placed at the bottom of
Each mesocosm was irrigated regularly by artificial behavior and the the chamber and inflowing air was passed through 25 mL of 0.01 M
water depth was maintained the same (i.e., 3 – 5 cm) to avoid the in­ H2SO4 to remove NH3 contamination from air. The air without NH3 was
fluence of water evaporation on the experimental results. extracted by the pump and flowed upward from the bottom of the
chamber to flow the gas generated in the chamber into the NH3 absor­
bent. During the NH3 collection, bubbles are always generated in the
2.3. Measurement of NH3 volatilization
safety bottle and the NH3 absorbent, which indicated that the chamber
has good sealing. The pH of the NH3 absorbent was adjusted to 6 – 7, and
NH3 volatilization was measured using a modified intermittent air-
the NH3 concentration was then measured on the automatic chemical
flow enclosure method (Fig. 1a and Fig. 1b). Sampling was conducted
analyzer (Mode Smartchem 200, Italy) by salicylate-sodium hypochlo­
from 10:30 a.m. to 2:30 p.m. local standard time on sampling days
rite spectrophotometric method. The calculation of NH3 volatilization
during the rice-growing season. There are 34 sampling days in total,
flux rate is described in Text S2 of Supplementary Materials.
including Day 0, 1, 3, 5, 7, 10 in BF stage, Day 0, 1, 3, 5, 7, 10, 15, 20, 25,
30, 37, 44, 51 in TF stage, and Day 0, 1, 3, 5, 7, 10, 15, 20, 25, 31, 38, 45,
52, 59, 61 in PIF stage. The size of the optically-clear static chamber was 2.4. Measurement of GHG emissions
450 mm length × 330 mm width × 700 mm height, and the detailed
description and operation of the chamber are presented in Text S1 of Fluxes of CH4, CO2 and N2O were simultaneously determined using
Supplementary Materials. The air tightness of all chambers was sys­ the closed static chamber-gas chromatograph method as detailed in our
tematically checked before the experiment, and the balance of air flow previous study (Chen et al. 2020b). Based on the optically-clear static
in and out of the chamber during NH3 collection was also determined. chamber for collecting NH3 volatilization, the outside of the chamber

Fig. 1. A diagram and real pictures of the experimental mesocosms with chambers (a and b, NH3 volatilization; c and d, GHG emissions). ① NH3 absorbent (500 mL
of 0.01 M H2SO4); ② Pump; ③ Cleaning bottle (25 mL of 0.01 M H2SO4); ④ Fan; ⑤ GHGs collection port; ⑥ Thermometer; ⑦ Pneumatic horizontal pipe.

3
X. Wang et al. Geoderma 433 (2023) 116460

was covered with shading cloth, and the gas collecting hole was replaced satisfied, the Greenhouse-Geisser correction was used to correct the
to make it a closed static chamber (Fig. 1c and Fig. 1d). The Chamber can degrees of freedom of the F-distribution. The significance of the differ­
avoid the respiratory effect of sunlight on plants which affects the ences in gas flux rates, microbial relative abundance, functional N-
collection of GHG, especially CO2. Gas samples were taken at 0, 20, 40 cycling gene abundance among treatments or days during the rice-
min after the chambers were placed on pre-fixed polyvinyl chloride growing season were assessed using the one-way ANOVA. The homo­
frames between 08:30 a.m. and 10:00 a.m. local standard time on geneity of variances was tested by using the Levene’s test. Multiple
sampling days of each fertilization during rice-growing season (same as comparisons of mean values were performed using the least significant
NH3 volatilization). Gas samples were collected from the chamber using difference (LSD). In all analyses, a p-value less than or equal to 0.05 was
the 100 mL vacuum sampling bags and were transported to the labo­ considered significant.
ratory for analysis by the gas chromatograph (GC-2010 Plus, Shimadzu
Corporation, Japan) within a few hours. The air temperature inside the 3. Results
chamber was monitored before and after gas extraction. The calcula­
tions of GHG flux rate and GWP are presented in Text S3 of Supple­ 3.1. NH3 volatilization
mentary Materials.
The volatilization rates and cumulative emissions of NH3 in the
2.5. Soil microbial abundance and diversity saline-alkali paddy ecosystem varied with different N-fertilizer treat­
ments but had a similar trend in time (Fig. 2). In the BF stage, the NH3
After the 137-day experiment (i.e., on October 14th), topsoil samples volatilization rates decreased gradually over time in all treatments. The
of all mesocosms with a depth of 0 – 10 cm were randomly collected in NH3 volatilization rates in all treatments with N-fertilizers (i.e., U, UI,
the rice rhizosphere by the quincunx sampling method with sterilized OCF and CSF) were significantly (p < 0.05) higher than CK on Day 7 of
50 mL centrifuge tubes and stored at − 20 ℃ (Cheng et al. 2007). All the BF stage (i.e., June 6th). The cumulative NH3 emissions by the end of
topsoil samples were sent to Sangon Biotech Co., ltd. (China) for sys­ BF stage were OCF > U > CSF > UI > CK, therein, a statistically sig­
tematic determination of microbial community structure. The total nificant (p < 0.05) difference was observed between OCF (36.76 ± 16.3
genome DNA extraction of each soil sample was performed using an E.Z. mg N m− 2) and CK (12.16 ± 2.07 mg N m− 2) treatments.
N.A. Soil DNA Kit (Omega Bio-tek, Inc., USA), following the standard In the TF stage, the NH3 volatilization rates increased from Day 1 (i.
protocol of manufacturer’s instructions. The concentration of the DNA e., June 14th) to Day 3 (i.e., June 16th), followed by a rapid decrease
was measured by using a Qubit 2.0 Fluorometer (Life Technologies, from Day 3 to Day 15 (i.e., June 28th), and then remained stable until
Carlsbad, CA, USA) to acquire adequate amounts of high-quality the end of TF stage for all treatments with N-fertilizers (Fig. 2a). The
genomic DNA. The polymerase chain reaction (PCR) amplification and highest peak value was observed in CSF treatment on Day 3 (1.89 ±
16S gene library construction, quantification, and sequencing followed 0.46 mg m-2h− 1) after the TF was applied, and its NH3 volatilization rate
the method described in our previous study (Wang et al. 2021c). The was significantly (p < 0.05) higher than all the other treatments. The
microbial communities were analyzed by high-throughput sequencing cumulative NH3 emissions by the end of TF stage (i.e., on Day 51 of the
(HTS) technology using the RDP database (https://rdp.cme.msu. TF stage, Aug. 3rd) were CSF > OCF > U > UI > CK (Fig. 2b). Therein, a
edu/misc/resources.jsp). Bacterial raw sequence reads of all topsoil statistically significant (p < 0.05) difference was observed between CK
samples were deposited in the National Center for Biotechnology In­ and all the other four treatments with N-fertilizers.
formation (NCBI) Sequence Read Archive (SRA) database under acces­ As described in Fig. 2a, from Day 7 to Day 15 of the PIF stage, the
sion number PRJNA946691. NH3 volatilization rates of UI treatment were significantly (p < 0.05)
higher than all other treatments. By the end of PIF stage (i.e., on Day 66
2.6. Abundance of functionals N-cycling genes of the PIF stage), the cumulative NH3 emission of UI treatment exceeded
that of OCF and CSF, respectively, and the cumulative NH3 emissions
The abundance of functional N-cycling genes in topsoil samples was were in the order UI > CSF > OCF > U > CK (Fig. 2b). Therein, the
quantified by real-time PCR (qPCR), which was also entrusted to Sangon cumulative NH3 emissions in UI (366.60 ± 55.47 mg N m− 2), OCF
Biotech Co., ltd. (China). The total soil genomic DNA was extracted (352.63 ± 31.22 mg N m− 2) and CSF (365.05 ± 40.78 mg N m− 2)
using the SanPrep Column Plasmid Mini-Preps Kit (Sangon Biotech Co. treatments were significantly (p < 0.05) higher than U (272.92 ± 30.09
ltd., Shanghai, China). The expression of the functional genes was mg N m− 2) and CK (154.86 ± 20.08 mg N m− 2) treatments, respectively.
analyzed by qPCR, and the PCR primers of each gene are shown in
Table S1 of Supplementary Materials. The qPCR procedure of each soil 3.2. GHG emissions
sample was carried out with a 10 μL mixture, including 1.0 μL of the
template DNA with tenfold serial dilutions, 5 μL of the SybrGreen qPCR The CH4 emission fluxes in all treatments increased from Day 0 of BF
Master Mix (2X), 0.2 μL of the F primer (10 μM), 0.2 μL of the R primer stage to Day 30 of TF stage (except for OCF, Day 25 of TF stage) and
(10 μM) and 3.6 μL of ddH2O. The qPCR was performed in the Light­ decreased thereafter to a stable level until Day 37 of TF stage (Fig. 3a).
cycler480 type II instrument (Roche Molecular Biochemicals, Man­ By the end of BF stage, the cumulative CH4 emission of OCF treatment
nheim, Germany) using the following program: one cycle of was higher than all the other treatments, therein, a statistically signifi­
denaturation at 95 ◦ C for 3 min, and 45 cycles of melting at 95 ◦ C for 5 s cant (p < 0.05) difference was observed between OCF (31.32 ± 11.80
and annealing at 60 ◦ C for 30 s. mg m− 2) and U (15.15 ± 0.45 mg m− 2) treatments. The cumulative CH4
emissions by the end of TF stage were OCF > CK > U > CSF > UI
2.7. Statistical analysis (Fig. 3b). After the entire rice-growing season (i.e., on Day 66 of PIF
stage), the cumulative CH4 emissions were CK > OCF > CSF > UI > U.
The statistical analyses and graphical interpretations of all experi­ The CO2 emission fluxes of all treatments increased from Day 0 of BF
mental data were performed by SPSS statistics 22.0 software (IBM stage to Day 44 of TF stage (i.e., from May 30th to July 27th), and then
Corporation, NY, USA) and Origin 9.1 software (OriginLab Corporation, decreased until Day 51 of TF stage (i.e., Aug. 3rd) (Fig. 3c). The CO2
USA) for Windows system. The data in all figures were presented as emission fluxes decreased from Day 0 to Day 5 of PIF stage (i.e., from
means ± standard deviation. A general linear model using a repeated Aug. 9th to 14th), increased from Day 10 of PIF stage (i.e., Aug. 19th)
measure analysis of variance (ANOVA) with Mauchly’s test of sphericity and gradually decreased to a stable level until Day 52 of TF stage (i.e.,
was used to determine the effect of gas (i.e., NH3, CH4, CO2 and N2O) Sep. 30th). After the entire rice-growing season, the cumulative CO2
flux rates and N-fertilizer types. When the spherical test was not emissions were CSF > U > OCF > CK > UI (Fig. 3d). In particular, the

4
X. Wang et al. Geoderma 433 (2023) 116460

Fig. 2. The volatilization rates (a) and cumulative

emissions (b) of NH3 in different N-fertilizer treat­
ments during the rice-growing season. BF: basal fer­
tilizer; TF: tillering fertilizer; PIF: panicle initiation
fertilizer. CK: treatment without N-Fertilizer; U:
treatment of urea as BF; UI: treatment of urea with
1% NBPT and 1% DMPP as BF; OCF: treatment of
organic–inorganic compound fertilizer as BF; CSF:
treatment of C-based slow-release fertilizer as BF.
Values represent the mean of three replicates and
error bars represent standard deviations (hereinafter
inclusive).

cumulative CO2 emission of UI treatment (1.34 ± 0.40 kg m− 2) biostatistics analysis, and the mean number of OTUs for all samples was
exhibited a significant (p < 0.05) difference with U (1.95 ± 0.08 kg m− 2) 2590 ± 694. The lowest and highest mean OTUs were observed in UI
and CSF (1.98 ± 0.31 kg m− 2), respectively. and OCF treatments, with the values of 2108 ± 164 and 3283 ± 395,
The N2O emission fluxes of all treatments exhibited a fluctuating respectively. The OTUs were analyzed for each sample with a sequence
trend (Fig. 3e). By the end of BF stage, the cumulative N2O emissions in similarity cut-off value of 97 %. The regularity of the Chao 1 and Ace
most N-fertilizer treatments (except for OCF) were higher than CK. The indexes, which were used to indicate the community richness in samples
cumulative N2O emissions by the end of TF stage were CSF > U > OCF > (Zhang et al. 2015), in all treatments was OCF > U > CSF > CK > UI. The
UI > CK (Fig. 3f). The cumulative N2O emissions of U (69.39 ± 15.79 Shannon and Simpson indexes were used to estimate the community
mg m− 2) and CSF (78.46 ± 12.71 mg m− 2) treatments were significantly diversity in samples (Classen et al. 2007). Both the maximum Shannon
(p < 0.05) higher than CK (21.49 ± 5.81 mg m− 2) and UI (29.65 ± 2.04 index (6.04 ± 0.50) and the minimum Simpson index (0.0161 ±
mg m− 2), respectively. By the end of the rice-growing season, the cu­ 0.0114) were observed in OCF treatment. Meanwhile, both the mini­
mulative N2O emissions were U > CSF > UI > CK > OCF (Fig. 3f). mum Shannon index (4.61 ± 0.25) and the maximum Simpson index
Therein, the cumulative N2O emissions in U (140.32 ± 49.33 mg m− 2) (0.0577 ± 0.0081) were observed in UI treatment. The Coverage indices
and CSF (88.84 ± 33.49 mg m− 2) treatments were significantly (p < of all samples were higher than 99.40 %, indicating that the given data
0.05) higher than CK (10.13 ± 22.73 mg m− 2) and OCF (-5.08 ± 36.11 were sufficient to reflect the distribution of microbial community in all
mg m− 2), respectively. Meanwhile, the cumulative N2O emission in UI treatments.
treatment (36.29 ± 18.84 mg m− 2) was significantly (p < 0.05) reduced The microbial community structures at phylum, class and genus
compared with U. The GWP values by the end of the entire rice-growing levels in all topsoil samples of different N-fertilizer treatments are pre­
season were CSF > U > OCF > CK > UI (Fig. 4). Therein, the GWP values sented in Fig. 5. As illustrated in Fig. 5a, the 5 dominant phyla in all
in CSF (2.06 × 104 ± 0.31 × 104 kg CO2-eq ha− 1) and U (2.04 × 104 ± treatments were the same, which were Proteobacteria (43.42 – 61.44 %),
0.08 × 104 kg CO2-eq ha− 1) treatments were significantly (p < 0.05) Firmicutes (5.54 – 13.02 %), Bacteroidetes (6.80 – 16.56 %), Chloroflexi
higher than UI (1.39 × 104 ± 0.39 × 104 kg CO2-eq ha− 1), respectively. (4.65 – 9.30 %) and Acidobacteria (1.48 – 4.49 %), respectively. Espe­
cially, the relative abundance of phylum Proteobacteria was significantly
(p < 0.05) higher than other phyla for all treatments. The maximum of
3.3. Microbial abundance, diversity and community structure
phylum Proteobacteria abundance was observed in CK treatment, while
the minimum was observed in OCF treatment. At class level, the species
As shown in Table 1, a total of 1,201,110 effective sequence reads
of microbial communities were basically the same among all treatments,
were obtained in topsoil samples of all treatments, and the mean
but different relative abundances of microbial communities were
sequence reads for all samples was 80074 ± 18507. These sequences
observed (Fig. 5b). Class Gammaproteobacteria exhibited the highest
produced 38843 of operational taxonomic units (OTUs) based on

5
X. Wang et al. Geoderma 433 (2023) 116460

Fig. 3. The GHG emission fluxes (a: CH4; c: CO2; e: N2O) and cumulative GHG (b: CH4; d: CO2; f: N2O) emissions in different N-fertilizer treatments during the rice-
growing season.

relative abundance among all classes for all treatments, and was 3.4. Abundance of functional genes
significantly (p < 0.05) higher than other classes. The relative abun­
dances of classes Gammaproteobacteria (18.95 %) and Bacilli (3.37 %) in As the preliminary DNA amplification results showed that pmoA gene
OCF treatment were significantly (p < 0.05) lower than those in all the could not be successfully amplified, the further detection of pmoA gene
other treatments, respectively. However, the relative abundance of class was halted. The abundance of functional genes involved in methano­
Alphaproteobacteria (8.58 %) in OCF treatment was significantly (p < genesis (i.e., mcrA gene), denitrification (i.e., nirS, nirK and nosZ genes)
0.05) increased compared with all the other treatments. At genus level, and the number of denitrifying bacteria involved in N2O reduction (i.e.,
the highest relative abundances of genera Escherichia-Shigella (16.77 %), (nirS + nirK) gene and (nirS + nirK)/nosZ ratio) in topsoil samples are
Acinetobacter (12.61 %), Vibrio (8.70 %), Sporosarcina (5.09 %), Klebsi­ presented in Fig. 6. For the methanogenesis, the highest abundance of
ella (2.80 %) and Geobacillus (2.48 %) were observed in UI treatment. mcrA gene was observed in topsoil of CK treatment (Fig. 6a), which was
The lowest relative abundances of genera Lysobacter (0.35 %), Ignavi­ significantly (p < 0.05) greater than UI, OCF and CSF, respectively. For
bacterium (0.38 %), Pontibacter (0.26 %), and Sphingomonas (0.71 %) the denitrification, the abundances of nirS, nirK and nosZ genes in topsoil
were also observed in UI treatment. Specifically, the relative abundance of OCF treatment were higher than CK. Therein, a statistically significant
of genera Escherichia-Shigella (4.99 %), Acinetobacter (3.96 %), Vibrio (p < 0.05) difference was observed in abundance of nirS gene between
(2.42 %) and Salmonella (1.51 %) in OCF treatment were significantly (p OCF (8.98 × 107 ± 1.27 × 107 copies g− 1 soil) and CK (3.81 × 107 ±
< 0.05) reduced compared with all the other treatments. The relative 0.68 × 107 copies g− 1 soil) treatments. Meanwhile, there was a signifi­
abundances of genus Sphingomonas (2.08 %) in OCF treatment was cant (p < 0.05) difference in the sum abundance of nirS and nirK genes
significantly (p < 0.05) higher than those in all the other treatments. between OCF and CK treatments (Fig. 6e). The ratios of (nirS + nirK)/
nosZ in topsoil were CSF > U > UI > OCF > CK (Fig. 6f). Therein, a
statistically significant (p < 0.05) difference was observed between CSF

6
X. Wang et al. Geoderma 433 (2023) 116460

Fig. 4. The GWP of different N-fertilizer treatments by the end of the entire rice-growing season. Columns containing different lowercase letters (a, b) indicate
significant difference among treatments at p = 0.05.

Table 1
Parameters related to microbial community richness and diversity in different N-fertilizer treatments (mean ± SD, n = 3).
Treatment Sequence reads OTUs Community richness Community diversity Coverage (%)
Chao 1 Ace Shannon Simpson

CK 70428 ± 10590 2274 ± 288 2513 ± 343 2498 ± 350 4.73 ± 0.38 0.0548 ± 0.0134 99.42 ± 0.20
U 92890 ± 21732 2886 ± 755 3178 ± 869 3151 ± 865 5.48 ± 0.86 0.0328 ± 0.0251 99.46 ± 0.33
UI 70189 ± 12819 2108 ± 164 2333 ± 235 2310 ± 239 4.61 ± 0.25 0.0577 ± 0.0081 99.46 ± 0.22
OCF 93711 ± 18773 3283 ± 395 3667 ± 452 3621 ± 440 6.04 ± 0.50 0.0161 ± 0.0114 99.40 ± 0.12
CSF 73152 ± 4744 2396 ± 797 2698 ± 982 2686 ± 967 4.92 ± 1.03 0.0490 ± 0.0310 99.41 ± 0.24

Note: OTUs: operational taxonomic units.

and all the other four treatments. 2016). Most studies use small-sized sampling chambers to collect NH3
volatilization, which cannot accommodate rice growing to maturity, so
4. Discussion NH3 volatilization in paddy soil cannot represent the complete paddy
ecosystems (Chu et al., 2020, Guo et al., 2021). The size of the sampling
4.1. Effect of N-fertilizer types on NH3 volatilization in paddy ecosystems chamber in this study was larger than in these studies, which resulted in
more water vapors to absorb the NH3 volatilized from paddy ecosys­
The N-fertilizer applied to soil is transformed into NH3 through a tems. Meanwhile, the larger size of the sampling chamber increased the
variety of enzymes (e.g., urease) and microbial activities (e.g., surface area of the chamber and thus the risk of NH3 absorption to the
ammonia-oxidizing archaea and bacteria) (Feng et al. 2021). In this chamber surface and absorption to condense water forming on the
study, the NH3 volatilization in actual environments of the paddy chamber surface. In addition, it is necessary to ensure that the inflow
ecosystem including plants (i.e., rice) was measured with reference to and outflow of gas are consistent when using a large-size sampling
the NH3 volatilization collection method of constructed wetlands (Du chamber to collect NH3, without NH3 leaking out. Otherwise, there is an
et al. 2018, Li et al. 2019). Although the NH3 volatilization flux and increased risk of underestimation of NH3 collection. Our results were
accumulation obtained in this study were lower than those in paddy soil consistent with the NH3 flux and accumulation evaluated by Li et al.
tested by other researchers (Chu et al. 2020, Guo et al. 2021), the overall using a sampling chamber with a similar size to our experiment in a
trend of NH3 volatilization flux was consistent with other studies paddy ecosystem (Li et al. 2019). Thus, even though the NH3 volatili­
(Fig. 2a). The main reasons for the relatively lower NH3 volatilization zation is likely underestimated by using the sampling chambers that can
flux observed in the paddy ecosystem in this study are as follows: (1) the accommodate rice plants, it does not affect the comparison of NH3
BF was spread before tillage to mix with soil at a depth of more than 6 cm volatilization of different N-fertilizers in saline-alkali paddy ecosystems.
by the deep fertilizer application method, which can increase the contact (3) As an indispensable part of the paddy ecosystem, rice can absorb NH3
between N-fertilizer particles and paddy soil. This procedure reduced for transforming N nutrients required for growth (Hayashi et al. 2008),
the rate of N-fertilizer hydrolysis, increased the soil absorption of NH+ 4, thus reducing the NH3 concentration in the surrounding atmosphere. (4)
and extended the effective period of N to achieve slow-release of N- Temperature, as an important environmental factor, has a positive
fertilizers (Liu et al. 2015, Yao et al. 2018). (2) NH3 volatilized from soil correlation with NH3 volatilization (Wang et al. 2021a). On most
is easily soluble in surface water and forms ammonia-N (NH+ 4 -N) (Hua experimental days of this study, the temperature was lower than in
et al. 2019, Zhao et al. 2016), and NH3 is easily absorbed by water vapor Southern China (e.g., Hunan and Hubei Province) (Fig. S1 of Supple­
due to the rapid rate of NH3 volatilization (Yang et al. 2020, Zhao et al. mentary Materials) where most studies on NH3 volatilization from

7
X. Wang et al. Geoderma 433 (2023) 116460

Fig. 5. Microbial composition of paddy topsoil at phylum (a), class (b) and genus (c) levels in different N-fertilizer treatments.

paddy fields were conducted (Wang et al. 2021a, Zhong et al. 2021). In comparing the actual NH3 volatilization under different N-fertilizer
summary, the method of collecting NH3 volatilization adopted in this applications with the same underestimated risk.
study, including topsoil, surface water, crops and water vapors, is closer In this study, different N-fertilizers were used as BF, mixed with
to the actual environment of the paddy ecosystem, thus reasonably paddy soil for deep application, and urea was used as TF and PIF in all N-

8
X. Wang et al. Geoderma 433 (2023) 116460

Fig. 6. Variations of the abundance of mcrA (a), nirS (b), nirK (c), nosZ (d), nirS + nirK (e) genes and relative abundance of (nirS + nirK)/nosZ (f) in topsoil of saline-
alkali paddy ecosystem with different N-fertilizer treatments. The central lines indicate median value in the box plots.

fertilizer treatments. The cumulative NH3 emissions in paddy meso­ BF stage. Additionally, nitrification inhibitors (e.g., DMPP), as a N-fer­
cosms by the end of BF stage were OCF > U > CSF > UI > CK, but the tilizer synergist, can inhibit the activities of nitrifying bacteria and ni­
cumulative NH3 emissions during the entre rice-growing season were UI trite bacteria (e.g., genera Ignavibacterium and Sphingomonas in this
> CSF > OCF > U > CK (Fig. 2b). For OCF treatment, the application of study) after being applied into soil (Fig. 5c), thereby inhibiting the
organic–inorganic compound fertilizer increased the organic matter conversion of NH+ 4 into NOx (e.g., N2O) in soil (Bal Krishna et al. 2013, Li
content of saline-alkali paddy soil. The addition of organic matter can et al. 2020, Wang et al. 2021b). Urease inhibitors (e.g., NBPT) with
enhance urease activity in soil and further promote the NH3 volatiliza­ hydrolase activity can reduce the hydrolysis rate of urea after applica­
tion (Matsushima et al. 2009), thereby increasing more NH3 volatiliza­ tion into soil by inhibiting the hydrolysis of C–N bonds, thereby
tion in paddy mesocosms. For CSF treatment, the application of C-based extending and/or controlling the fertilizer efficiency (Cantarella et al.
slow-release fertilizer only in the BF stage reduced the NH3 volatilization 2018). By the end of TF stage, in UI treatment, the N conversion process
in paddy mesocosms compared with U treatment (Fig. 2b). During the was inhibited by both urease and nitrification inhibitors, which delayed
entire rice-growing season, although rice growth was improved (see the conversion and loss of N in paddy soil, thus reducing the cumulative
Fig. S2 of Supplementary Materials), cumulative NH3 volatilization in NH3 emission compared with all the other N-fertilizer treatments
CSF treatment was significantly (p < 0.05) higher than U treatment (Fig. 2b). However, the content of urease inhibitor might be decreased in
(Fig. 2b). The reason may be that C-based slow-release fertilizer when UI treatment at the PIF stage, which failed to exert its function, resulting
just applied to paddy soil can slow or control the nutrient release in more NH3 accumulation compared with all the other N-fertilizer
compared with traditional urea (Dong et al. 2020). However, from the treatments (Fig. 2b). In general, compared with traditional urea, C-
view of the entire rice-growing season, the addition of biochar from C- based slow-release fertilizer reduced the NH3 accumulation in a saline-
based slow-release fertilizer accelerated the NH3 volatilization rates and alkali paddy ecosystem in the BF stage, and urea with inhibitors can
thereby causing higher cumulative NH3 volatilization losses (Feng et al. effectively control NH3 volatilization in the BF and TF stages.
2017). Thus, the application of C-based slow-release fertilizer only
controlled NH3 volatilization in a saline-alkali paddy ecosystem in the

9
X. Wang et al. Geoderma 433 (2023) 116460

4.2. Effect of N-fertilizer types on GHG emissions in paddy ecosystems emission in saline-alkali paddy fields.

4.2.1. CH4 emission 4.2.3. N2O emission

Paddy ecosystem is one of the important sources of GHG emissions N2O is a long-lived potent GHG with the greatest warming potential,
(Jiang et al. 2021, Song et al. 2021). On a global scale, about 10 % of which has the most important impact on the GWP and emission intensity
total CH4 emissions are contributed by rice cultivation (Cao et al. 2021). of farmland (Braker and Conrad 2011). N2O emission is mainly pro­
The production of CH4 in paddy fields mainly comes from the SOM duced by microbial nitrification and denitrification in paddy soil, and is
decomposition, driven by a series of complex soil microbial processes the result of abiotic factors regulating the N2O generation and reduction
(Zhao et al. 2019). In an anaerobic environment, organic matter (OM, e. process of microorganisms (Bhattacharyya et al. 2013, Braker and
g., SOM, root exudates, dead crop roots and residues, dead soil animals Conrad 2011). The N2O emission from paddy fields largely depends on
and microorganisms, applied organic fertilizer, etc.) is gradually the input level of N-fertilizer (Montoya et al. 2021). Unreasonable or
degraded into organic acids, alcohols, CO2 and other small molecule excessive application of N-fertilizer can cause direct emission of N2O in
compounds by anaerobic bacteria, and the molecule compounds are paddy soil (Liu et al. 2022). After being applied to the paddy soil, N-
then converted into CH4 by methanogens (Ma et al. 2010, Zhao et al. fertilizer is quickly converted into NH+ 4 , and is further converted into
2019). Rice growth is promoted by the N-fertilizer application, allowing NO–3 by nitrification (Han et al. 2020). The direct contribution of nitri­
more root exudates and rice litter to enter the soil, providing more OM fication to N2O emission is relatively low, and it was lower than 1 % of
for CH4 production (Zhao et al. 2019). For OCF treatment in this study, the total amount of ammonia-N oxidation in most cases (Chen et al.
the organic–inorganic compound fertilizer is rich in OM, which is mainly 2013). Although N2O is not the intermediate product of nitrification
the fermented and decomposed OM from animal and plant residues. As process, its production is closely related to the intermediate products (i.
presented in Fig. 3b, the cumulative CH4 emission of OCF treatment e., hydroxylamine (NH2OH) and nitrite (NO–2)) and the product in soil
after BF stage was higher than in other treatments. Simultaneously, nitrification (i.e., NO–3). The combination of two or more inhibitors can
more developed rice plants also provide a better release channel for CH4. effectively alleviate N2O emission in paddy fields when applied at a
However, with the passage of time, the growth of rice plants after suitable dosage (Corrochano-Monsalve et al. 2021, Recio et al. 2020).
fertilization was promoted, thus increasing the number of aeration tis­ Both NBPT and DMPP in UI treatment were used to regulate the activ­
sues. Oxygen in the atmosphere enters the roots through the aeration ities of enzymes and microorganisms in paddy soil (Fig. 5), and prolong
tissue of rice plants, forming aerobic micro-regions to oxidize CH4 in the the residual time of N-fertilizer (e.g., urea) (Cantarella et al. 2018, Li
rhizosphere of rice (Jiang et al. 2017). After the entire rice-growing et al. 2020). Especially, DMPP, as the nitrification inhibitor in UI
season, the cumulative CH4 emission in CK treatment was higher than treatment, was used to inhibit the nitrification process of NH+ 4 into NO3
–

those in all N-fertilizer treatments (Fig. 3b). Moreover, for all N-fertilizer by ammonia-oxidizing bacteria after the urea application in paddy soil
treatments, the possible reason for the reduction of CH4 emission may be (Li et al. 2020, Liu et al. 2022). The reduction of NO–3, which is the
the decrease of methanogenic bacteria community diversity, abundance nitrification product, leads to the inhibition of its denitrification to N2O
and activity compared with CK treatment (Han et al. 2013). The growth (Wu et al. 2022).Thus, the addition of urease-nitrification inhibitors (i.
of methanogens can be inhibited by the application of N-fertilizers in e., NBPT and DMPP) effectively controlled the N2O emission in saline-
paddy fields (Ma et al. 2007, Wang et al. 2020). For example, the relative alkali paddy fields, causing significantly (p < 0.05) lower cumulative
abundance of genus Acinetobacter, which is capable of producing CH4 N2O emission and GWP in UI treatment than U (Fig. 3f and Fig. 4).
(Iltchenco et al. 2021), was significantly (p < 0.05) lower in most N- The abundance of N functional microorganisms in soil is closely
fertilizer treatments (except for UI) than CK (Fig. 5c). At the molecular related to soil nutrient transformation and GHG emissions. For OCF
level, the last step of CH4 production is controlled by the mcrA gene of treatment in this study, the OTUs, Chao 1, Ace, Shannon indexes were
methanogens, which are the only source of CH4 production in paddy higher than those of all other treatments (Table 1), indicating that the
fields (Watanabe et al. 2007). In this study, the mcrA gene abundances in application of organic–inorganic compound fertilizer improved the mi­
topsoil of all the N-fertilizer treatments were decreased compared with crobial richness and diversity in saline-alkali paddy fields. Especially,
CK (Fig. 6a), which also verified the reason mentioned above. the application of organic–inorganic compound fertilizer can effectively
improve the abundance of microorganisms involved in N cycle in saline-
4.2.2. CO2 emission alkali paddy fields (Fig. 5), e.g., nitrifying bacteria (genus Sphingomonas)
CO2 emission in a paddy ecosystem is mainly caused by soil respi­ and nitrogen-fixing bacteria (class Alphaproteobacteria), etc. (Bal Krishna
ration (Iqbal et al. 2009, Xu and Shang 2016). The application of N- et al. 2013, He et al. 2018, Tsoy et al. 2016). Denitrification, as a mi­
fertilizer changed the biomass and activity of rice roots, the number and crobial respiration process, can restore NO–3 to N2 and produce N2O that
activity of microorganisms and soil chemical properties, which caused destroys the ozone layer under flooded paddy soil (Ishii et al. 2011). The
the change of soil respiration (Bhattacharyya et al. 2013). Fertilization application of N-fertilizer stimulated the abundance of denitrifying
promotes rice growth, accelerates growth of aboveground and below­ bacteria in saline-alkali paddy fields and increased the nirS gene abun­
ground biomass, and increases autotrophic respiration of rice. Fertil­ dance of denitrifying bacteria in paddy soil (Fig. 6b). Therein, the nirS
ization also can increase the amount of decomposable SOM, which gene in topsoil of OCF treatment was significantly (p < 0.05) higher than
provides more substrates for microbial activity and increases the mi­ CK. The nosZ is a common gene encoding N2O reductase that reduces
crobial activity, thereby promoting the CO2 emission of soil respiration N2O to N2 (Chen et al. 2020a), and its abundance in the OCF treatment
(Iqbal et al. 2009). In addition, the application of N-fertilizers can reduce was also increased compared with CK (Fig. 6d). The results showed that
the C/N ratio of soil, enhance the activity of soil microorganisms, and the application of organic–inorganic compound fertilizer had a stronger
improve the CO2 emission (Liu et al. 2020). In this study, after the stimulating effect on the abundance of denitrifying bacteria in saline-
application of most N-fertilizers (except for UI), the richness and di­ alkali paddy fields. The (nirS + nirK)/nosZ ratio could be a good indi­
versity of the soil microbial community in saline-alkali paddy fields was cator for predicting N2O emission (Chen et al. 2020a). The (nirS + nirK)/
increased (Table 1). However, for UI treatment, the addition of NBPT nosZ ratio is greater than 1:1, indicating that there are more denitrifying
and DMPP reduced the number of soil microorganisms in saline-alkali bacteria involved in N2O reduction (Liu et al. 2018). For OCF treatment
paddy fields, which was verified by the microbial data (Table 1). The in this study, the (nirS + nirK)/nosZ ratio was lower than all the other N-
cumulative CO2 emissions in UI treatment was decreased compared to fertilizer treatments (Fig. 6f), and the cumulative N2O emission was
all the other treatments, and it exhibited a significant (p < 0.05) dif­ significantly (p < 0.05) reduced (Fig. 3f). Furthermore, the sum of cu­
ference with CSF and U treatments (Fig. 3d). Consequently, urea with mulative N2O and NH3 emissions in OCF treatment was lower than all
inhibitors (i.e., NBPT and DMPP) is beneficial for controlling CO2 the other N-fertilizer treatments (see Fig. S3 of Supplementary

10
X. Wang et al. Geoderma 433 (2023) 116460

Materials). Therefore, organic–inorganic compound fertilizer can N dynamics in relation to CH4 and N2O emissions from tropical flooded rice (Oryza
sativa L.). Sci. Total Environ. 461–462 (7), 601–611.
effectively control N2O emission from saline-alkali paddy fields, and
Braker, G., Conrad, R., 2011. Diversity, structure, and size of N2O-producing microbial
reduce N loss in gaseous form. In future studies, we recommend to communities in soils–what matters for their functioning? Adv. Appl. Microbiol. 75,
expand the scale of experiment and consider the pathways of N loss via 33–70.
rainstorm runoff and infiltration to further explore the response of N loss Cantarella, H., Otto, R., Soares, J.R., Silva, A.G.B., 2018. Agronomic efficiency of NBPT
as a urease inhibitor: A review. J. Adv. Res. 13, 19–27.
and GHG emissions to N-fertilizer types in saline-alkali paddy fields. Cao, Y., Shan, Y., Wu, P., Zhang, P., Zhang, Z., Zhao, F., Zhu, T., 2021. Mitigating the
global warming potential of rice paddy fields by straw and straw-derived biochar
5. Conclusions amendments. Geoderma 396, 115081.
Chen, H., Li, X., Hu, F., Wei, S., 2013. Soil nitrous oxide emissions following crop residue
addition: a meta-analysis. Global Change Biol. 19 (10), 2956–2964.
This study investigated the effect of different N-fertilizers on NH3 Chen, X., Zhu, H., Yan, B., Shutes, B., Tian, L., Wen, H., 2020a. Optimal influent COD/N
volatilization, GHG emissions, the structure and diversity of the micro­ ratio for obtaining low GHG emissions and high pollutant removal efficiency in
constructed wetlands. J. Clean. Prod. 267, 122003.
bial community and functional gene abundance in a saline-alkali paddy Chen, X., Zhu, H., Yan, B., Shutes, B., Xing, D., Banuelos, G., Cheng, R., Wang, X., 2020b.
ecosystem. The cumulative NH3 emissions of CSF and UI treatments in Greenhouse gas emissions and wastewater treatment performance by three plant
the BF stage were lower than U and OCF, respectively, while the cu­ species in subsurface flow constructed wetland mesocosms. Chemosphere 239,
124795.
mulative NH3 emissions of UI, OCF and CSF treatments were signifi­ Cheng, J.L., Shi, Z., Zhu, Y.W., 2007. Assessment and mapping of environmental quality
cantly (p < 0.05) higher compared with U for the entire rice-growing in agricultural soils of Zhejiang Province, China. J Environ. Sci. 29 (1), 50–54.
season. After the entire rice-growing season, the cumulative CH4 emis­ Chi, C.M., Zhao, C.W., Sun, X.J., Wang, Z.C., 2012. Reclamation of saline-sodic soil
properties and improvement of rice (Oriza sativa L.) growth and yield using
sions in all N-fertilizer treatments were lower than CK. The cumulative
desulfurized gypsum in the west of Songnen Plain, northeast China. Geoderma
CO2 emissions were U > CSF > OCF > CK > UI, and the cumulative N2O 187–188, 24–30.
emissions were U > CSF > UI > CK > OCF. Compared with all the other Chu, Q., Xue, L., Cheng, Y., Liu, Y., Feng, Y., Yu, S., Meng, L., Pan, G., Hou, P., Duan, J.,
N-fertilizers, the application of organic–inorganic compound fertilizer Yang, L., 2020. Microalgae-derived hydrochar application on rice paddy soil: Higher
rice yield but increased gaseous nitrogen loss. Sci. Total Environ. 717, 137127.
can effectively improve the relative abundances of nitrifying bacteria Classen, A.T., Overby, S.T., Hart, S.C., Koch, G.W., Whitham, T.G., 2007. Season
(genus Sphingomonas) and nitrogen-fixing bacteria (class Alphaproteo­ mediates herbivore effects on litter and soil microbial abundance and activity in a
bacteria), and reduce the (nirS + nirK)/nosZ ratio in topsoil, thereby semi-arid woodland. Plant Soil 295 (1–2), 217–227.
Corrochano-Monsalve, M., Bozal-Leorri, A., Sánchez, C., González-Murua, C., Estavillo, J.
controlling N2O emissions in saline-alkali paddy ecosystems. The GWP M., 2021. Joint application of urease and nitrification inhibitors to diminish gaseous
values were CSF > U > OCF > CK > UI, and the UI treatment value was nitrogen losses under different tillage systems. J. Clean. Prod. 289 (3), 125701.
significantly (p < 0.05) lower compared with U and CSF treatments, Dong, D., Wang, C., Van Zwieten, L., Wang, H., Jiang, P., Zhou, M., Wu, W., 2020. An
effective biochar-based slow-release fertilizer for reducing nitrogen loss in paddy
respectively. To summarize, the application of urea with urease- fields. J. Soil. Sediment. 20 (8), 3027–3040.
nitrification inhibitors (i.e., NBPT and DMPP) can control N loss via Du, L., Trinh, X., Chen, Q., Wang, C., Liu, S., Liu, P., Zhou, Q., Xu, D., Wu, Z., 2018. Effect
NH3 volatilization in saline-alkali paddy ecosystems in the BF stage. If of clinoptilolite on ammonia emissions in integrated vertical-flow constructed
wetlands (IVCWs) treating swine wastewater. Ecol. Eng. 122, 153–158.
considering the entire rice-growing season, urea has the potential to Feng, Y., Sun, H., Xue, L., Liu, Y., Gao, Q., Lu, K., Yang, L., 2017. Biochar applied at an
reduce NH3 volatilization. The application of urea with urease- appropriate rate can avoid increasing NH3 volatilization dramatically in rice paddy
nitrification inhibitors can effectively reduce GHG emissions in saline- soil. Chemosphere 168, 1277–1284.
Feng, Y., He, H., Li, D., He, S., Yang, B., Xue, L., Chu, Q., 2021. Biowaste hydrothermal
alkali paddy fields, thus reducing GWP, but resulting in higher NH3
carbonization aqueous product application in rice paddy: Focus on rice growth and
volatilization than urea. While organic–inorganic compound fertilizer ammonia volatilization. Chemosphere 277, 130233.
application mainly reduced N2O emission and N loss in gaseous form (i. Guo, J., Fan, J., Zhang, F., Yan, S., Zheng, J., Wu, Y., Li, J., Wang, Y., Sun, X., Liu, X.,
e., the sum of cumulative N2O and NH3 emissions), but did not affect Xiang, Y., Li, Z., 2021. Blending urea and slow-release nitrogen fertilizer increases
dryland maize yield and nitrogen use efficiency while mitigating ammonia
GWP. volatilization. Sci. Total Environ. 790, 148058.
Han, P., Wu, D., Sun, D., Zhao, M., Yin, G., 2020. N2O and NOy production by the
Declaration of Competing Interest comammox bacterium Nitrospira inopinata in comparison with canonical ammonia
oxidizers. Water Res. 190 (6), 116728.
Han, C., Zhong, W., Shen, W., Cai, Z., Liu, B., 2013. Transgenic Bt rice has adverse
The authors declare that they have no known competing financial impacts on CH4 flux and rhizospheric methanogenic archaeal and methanotrophic
interests or personal relationships that could have appeared to influence bacterial communities. Plant Soil 369 (1–2), 297–316.
Hassani, A., Azapagic, A., Shokri, N., 2020. Predicting long-term dynamics of soil salinity
the work reported in this paper. and sodicity on a global scale. Proc. Natl. Acad. Sci. USA 117 (52), 33017–33027.
Hayashi, K., Nishimura, S., Yagi, K., 2008. Ammonia volatilization from a paddy field
Data availability following applications of urea: rice plants are both an absorber and an emitter for
atmospheric ammonia. Sci. Total Environ. 390 (2–3), 485–494.
He, Q., Feng, C., Chen, N., Zhang, D., Hou, T., Dai, J., Hao, C., Mao, B., 2018.
Data will be made available on request. Characterizations of dissolved organic matter and bacterial community structures in
rice washing drainage (RWD)-based synthetic groundwater denitrification.
Chemosphere 215, 142–152.
Acknowledgments
Hua, L., Zhai, L., Liu, J., Guo, S., Li, W., Zhang, F., Fan, X., Liu, H., 2019. Characteristics
of nitrogen losses from a paddy irrigation-drainage unit system. Agr. Ecosyst.
This study was supported by the Excellent Youth Foundation of Jilin Environ. 285, 106629.
Iltchenco, J., Peruzzo, V., Eva Magrini, F., Marconatto, L., Paula Torres, A., Luiz Beal, L.,
Province, China (No. 20230101361JC), the CAS Interdisciplinary
Paesi, S., 2021. Microbiota profile in mesophilic biodigestion of sugarcane vinasse in
Innovation Team Project (No. JCTD-2020-14), and the Youth Innovation batch reactors. Water Sci. Technol. 84 (8), 2028–2039.
Promotion Association, CAS (No. 2017274 and No. Y2021068). Iqbal, J., Hu, R., Lin, S., Hatano, R., Feng, M., Lu, L., Ahamadou, B., Du, L., 2009. CO2
emission in a subtropical red paddy soil (Ultisol) as affected by straw and N-fertilizer
applications: A case study in Southern China. Agr. Ecosyst. Environ. 131 (3–4),
Appendix A. Supplementary data 292–302.
Ishii, S., Ikeda, S., Minamisawa, K., Senoo, K., 2011. Nitrogen cycling in rice paddy
Supplementary data to this article can be found online at https://doi. environments: Past achievements and future challenges. Microbes. Environ. 26 (4),
282–292.
org/10.1016/j.geoderma.2023.116460. Jeppesen, E., Beklioğlu, M., Özkan, K., Akyürek, Z., 2020. Salinization increase due to
climate change will have substantial negative effects on inland waters and
References freshwater resources: A call for multifaceted research at the local and global scale.
The Innovation 1 (2), 100030.
Jiang, Y., van Groenigen, K.J., Huang, S., Hungate, B.A., van Kessel, C., Hu, S., Zhang, J.,
Bal Krishna, K.C., Sathasivan, A., Ginige, M.P., 2013. Microbial community changes with
Wu, L., Yan, X., Wang, L., Chen, J., Hang, X., Zhang, Y., Horwath, W.R., Ye, R.,
decaying chloramine residuals in a lab-scale system. Water Res. 47 (13), 4666–4679.
Linquist, B.A., Song, Z., Zheng, C., Deng, A., Zhang, W., 2017. Higher yields and
Bhattacharyya, P., Roy, K.S., Neogi, S., Dash, P.K., Nayak, A.K., Mohanty, S., Baig, M.J.,
Sarkar, R.K., Rao, K.S., 2013. Impact of elevated CO2 and temperature on soil C and

11
X. Wang et al. Geoderma 433 (2023) 116460

lower methane emissions with new rice cultivars. Global Change Biol. 23 (11), Wang, B., Li, R., Wan, Y., Li, Y., Cai, W., Guo, C., Qin, X., Song, C., Wilkes, A., 2021a. Air
4728–4738. warming and CO2 enrichment cause more ammonia volatilization from rice paddies:
Jiang, Z., Yang, S., Pang, Q., Xu, Y., Chen, X., Sun, X., Qi, S., Yu, W., 2021. Biochar An OTC field study. Sci. Total Environ. 752, 142071.
improved soil health and mitigated greenhouse gas emission from controlled Wang, J., Liang, J., Sun, L., Shen, J., Wang, M., 2021b. Achieving reliable partial
irrigation paddy field: Insights into microbial diversity. J. Clean. Prod. 318, 128595. nitrification and anammox process using polyvinyl alcohol gel beads to treat low-
Li, H., Chen, X., Liu, C., Wang, S., Guo, B., Li, N., Jin, Y., Fu, Q., Liang, X., 2020. Effect of strength ammonia wastewater. Bioresour. Technol. 324, 124669.
various doses of 3,4-dimethylpyrazole phosphate on mineral nitrogen losses in two Wang, X., Zhu, H., Yan, B., Shutes, B., Bañuelos, G., Wen, H., 2020. Bioaugmented
paddy soils. J. Soil. Sediment. 20 (11), 3825–3834. constructed wetlands for denitrification of saline wastewater: A boost for both
Li, F., Feng, J., Zhou, X., Xu, C., Haissam Jijakli, M., Zhang, W., Fang, F., 2019. Impact of microorganisms and plants. Environ. Int. 138, 105628.
rice-fish/shrimp co-culture on the N2O emission and NH3 volatilization in intensive Wang, X., Zhu, H., Yan, B., Shutes, B., Banuelos, G., Cheng, R., 2021c. Response of the
aquaculture ponds. Sci. Total Environ. 655, 284–291. microbial community to salt stress and its stratified effect in constructed wetlands.
Li, Q., Yang, A., 2020. Comparative studies on seed germination of two rice genotypes Environ. Sci. Pollut. Res. 28 (14), 18089–18101.
with different tolerances to low temperature. Environ. Exp. Bot. 179, 104216. Watanabe, T., Kimura, M., Asakawa, S., 2007. Dynamics of methanogenic archaeal
Liu, T.Q., Fan, D.J., Zhang, X.X., Chen, J., Li, C.F., Cao, C.G., 2015. Deep placement of communities based on rRNA analysis and their relation to methanogenic activity in
nitrogen fertilizers reduces ammonia volatilization and increases nitrogen utilization Japanese paddy field soils. Soil Biol. Biochem. 39 (11), 2877–2887.
efficiency in no-tillage paddy fields in central China. Field Crop. Res. 184, 80–90. Wong, V.N.L., Greene, R.S.B., Dalal, R.C., Murphy, B.W., 2010. Soil carbon dynamics in
Liu, J., Li, C., Jing, J., Zhao, P., Luo, Z., Cao, M., Ma, Z., Jia, T., Chai, B., 2018. Ecological saline and sodic soils: A review. Soil Use Manage. 26 (1), 2–11.
patterns and adaptability of bacterial communities in alkaline copper mine drainage. Wu, K., Gong, P., Bai, W., Zhang, Z., Wei, Z., Yu, C., Song, Y., Xue, Y., Zhang, L., 2022.
Water Res. 133, 99–109. Effect of mixed inhibitor application on N2O production pathways in paddy soil.
Liu, Y., Shahbaz, M., Ge, T., Zhu, Z., Liu, S., Chen, L., Wu, X., Deng, Y., Lu, S., Wu, J., J. Soil. Sediment. 22 (7), 1913–1923.
2020. Effects of root exudate stoichiometry on CO2 emission from paddy soil. Eur. J. Xu, M., Shang, H., 2016. Contribution of soil respiration to the global carbon equation.
Soil Biol. 101, 103247. J. Plant Physiol. 203, 16–28.
Liu, C., Zhang, Y., Liu, H., Liu, X., Ren, D., Wang, L., Guan, D., Li, Z., Zhang, M., 2022. Yang, Y., Li, N., Ni, X., Yu, L., Yang, Y., Wang, Q., Liu, J., Ye, Y., Tao, L., Liu, B., Wu, Y.,
Fertilizer stabilizers reduce nitrous oxide emissions from agricultural soil by 2020. Combining deep flooding and slow-release urea to reduce ammonia emission
targeting microbial nitrogen transformations. Sci. Total Environ. 806, 151225. from rice fields. J. Clean. Prod. 244, 118745.
Ma, J., Li, X.L., Xu, H., Han, Y., Cai, Z.C., Yagi, K., 2007. Effects of nitrogen fertiliser and Yang, X., Shang, Q., Wu, P., Liu, J., Shen, Q., Guo, S., Xiong, Z., 2010. Methane emissions
wheat straw application on CH4 and N2O emissions from a paddy rice field. Aust. J. from double rice agriculture under long-term fertilizing systems in Hunan, China.
Soil Res. 45 (5), 359–367. Agr. Ecosyst. Environ. 137 (3–4), 308–316.
Ma, K.E., Qiu, Q., Lu, Y., 2010. Microbial mechanism for rice variety control on methane Yao, Y., Zhang, M., Tian, Y., Zhao, M., Zhang, B., Zhao, M., Zeng, K., Yin, B., 2018. Urea
emission from rice field soil. Global Change Biol. 16 (11), 3085–3095. deep placement for minimizing NH3 loss in an intensive rice cropping system. Field
Matsushima, M., Lim, S.S., Kwak, J.H., Park, H.J., Lee, S.I., Lee, D.S., Choi, W.J., 2009. Crop. Res. 218, 254–266.
Interactive effects of synthetic nitrogen fertilizer and composted manure on Yu, H., Wang, Z., Mao, D., Jia, M., Chang, S., Li, X., 2023. Spatiotemporal variations of
ammonia volatilization from soils. Plant Soil 325 (1–2), 187–196. soil salinization in China’s West Songnen Plain. Land Degrad. Dev. https://doi.org/
Montoya, M., Vallejo, A., Corrochano-Monsalve, M., Aguilera, E., Sanz-Cobena, A., 10.1002/ldr.4613.
Ginés, C., Guardia, G., 2021. Mitigation of yield-scaled nitrous oxide emissions and Yu, L.X., Yang, J.C., Bu, K., Liu, T., Jiao, Y., Li, G.S., Pu, L., M., Zhang, S.W., 2021.
global warming potential in an oilseed rape crop through N source management. Impacts of saline-alkali land improvement on regional climate: Process, mechanisms,
J. Environ. Manage. 288, 112304. and implications. Remote Sens. (Basel) 13 (17), 3407.
Pan, Y., Tian, S., Zhao, Y., Zhang, L., Zhu, X., Gao, J., Huang, W., Zhou, Y., Song, Y., Zhan, X., Adalibieke, W., Cui, X., Winiwarter, W., Reis, S., Zhang, L., Bai, Z., Wang, Q.,
Zhang, Q., Wang, Y., 2018. Identifying ammonia hotspots in China using a national Huang, W., Zhou, F., 2021. Improved estimates of ammonia emissions from global
observation network. Environ. Sci. Technol. 52, 3926–3934. croplands. Environ. Sci. Technol. 55 (2), 1329–1338.
Qadir, M., Oster, J.D., Schubert, S., Noble, A.D., Sahrawat, K.L., 2007. Phytoremediation Zhang, H.H., Huang, T.L., Chen, S.N., Yang, X., Lv, K., Sekar, R., 2015. Abundance and
of sodic and saline-sodic soils. Adv. Agron. 96, 197–247. diversity of bacteria in oxygen minimum drinking water reservoir sediments studied
Recio, J., Montoya, M., Ginés, C., Sanz-Cobena, A., Vallejo, A., Alvarez, J.M., 2020. Joint by quantitative PCR and pyrosequencing. Microb. Ecol. 69 (3), 618–629.
mitigation of NH3 and N2O emissions by using two synthetic inhibitors in an Zhang, Q., Pan, Y., He, Y., Zhao, Y., Zhu, L., Zhang, X., Xu, X., Ji, D., Gao, J., Tian, S.,
irrigated cropping soil. Geoderma 373, 114423. Gao, W., Wang, Y., 2019. Bias in ammonia emission inventory and implications on
Riya, S., Zhou, S., Watanabe, Y., Sagehashi, M., Terada, A., Hosomi, M., 2012. CH4 and emission control of nitrogen oxides over North China Plain. Atmos. Environ. 214,
N2O emissions from different varieties of forage rice (Oryza sativa L.) treating liquid 116869.
cattle waste. Sci. Total Environ. 419, 178–186. Zhang, S., Zhang, G., Wang, D., Liu, Q., Xu, M., 2021. Investigation into runoff nitrogen
Shan, L., He, Y., Chen, J., Huang, Q., Wang, H., 2015. Ammonia volatilization from a loss variations due to different crop residue retention modes and nitrogen fertilizer
Chinese cabbage field under different nitrogen treatments in the Taihu Lake Basin. rates in rice-wheat cropping systems. Agr. Water Manage. 247, 106729.
China. J. Environ. Sci. 38 (12), 14–23. Zhao, X., Yan, X., Xie, Y., Wang, S., Xing, G., Zhu, Z., 2016. Use of nitrogen isotope to
Song, K., Zhang, G., Yu, H., Huang, Q., Zhu, X., Wang, T., Xu, H., Lv, S., Ma, J., 2021. determine fertilizer- and soil-derived ammonia volatilization in a rice/wheat
Evaluation of methane and nitrous oxide emissions in a three-year case study on rotation system. J. Agric. Food Chem. 64 (15), 3017–3024.
single rice and ratoon rice paddy fields. J. Clean. Prod. 297, 126650. Zhao, X., Pu, C., Ma, S.T., Liu, S.L., Xue, J.F., Wang, X., Wang, Y.Q., Li, S.S., Lal, R.,
Sun, C., Chen, L., Liu, H.B., Zhu, H., Lü, M.Q., Chen, S.B., Wang, Y.W., Shen, Z.Y., 2021. Chen, F., Zhang, H.L., 2019. Management-induced greenhouse gases emission
New modeling framework for describing the pollutant transport and removal of mitigation in global rice production. Sci. Total Environ. 649, 1299–1306.
ditch-pond system in an agricultural catchment. Water Resour. Res. 57. Zhao, C., Zhang, H., Song, C., Zhu, J.-K., Shabala, S., 2020a. Mechanisms of plant
Sun, L., Wu, Z., Ma, Y., Liu, Y., Xiong, Z., 2018. Ammonia volatilization and atmospheric responses and adaptation to soil salinity. The Innovation 1 (1), 100017.
N deposition following straw and urea application from a rice-wheat rotation in Zhao, W., Zhou, Q., Tian, Z., Cui, Y., Liang, Y., Wang, H., 2020b. Apply biochar to
southeastern China. Atmos. Environ. 181, 97–105. ameliorate soda saline-alkali land, improve soil function and increase corn nutrient
Tsoy, O.V., Ravcheev, D.A., Jelena, Č., Gelfand, M.S., 2016. Nitrogen fixation and availability in the Songnen Plain. Sci. Total Environ. 722, 137428.
molecular oxygen: Comparative genomic reconstruction of transcription regulation Zhong, X., Zhou, X., Fei, J., Huang, Y., Wang, G., Kang, X., Hu, W., Zhang, H., Rong, X.,
in Alphaproteobacteria. Front. Microbiol. 7 (201), 1343. Peng, J., 2021. Reducing ammonia volatilization and increasing nitrogen use
efficiency in machine-transplanted rice with side-deep fertilization in a double-
cropping rice system in Southern China. Agr. Ecosyst. Environ. 306, 107183.

12