Estimation of nitrous oxide emissions from rice paddy fields using the DNDC model: a case study of South Korea Open Access

Nitrous oxide (N₂O) is well known greenhouse gas for its global warming potential of 298 times greater than carbon dioxide (CO₂) (IPCC 2007). N₂O is also known for its role in damaging the stratospheric ozone layer (Fan et al. 2020). The soil ecosystem accounts for 65% (6.8 Tg N₂O-N year⁻¹) of the total N₂O emitted to the atmosphere mainly due to nitrogen fertilizer application (IPCC 2007). Nitrogen is the most deficient nutrient in soils and essential for plant growth. Nitrogen fertilizer is extensively used to meet the plant need and food demand at large. However, 1–5% of N fertilizer applied is lost into the atmosphere as N₂O emissions (Tang et al. 2018; Liu et al. 2019a). Nitrification and denitrification in the soil are core pathways of N₂O emissions (Wei et al. 2017; Duan et al. 2019). Nitrification occurs under aerobic conditions in which primary nitrifiers (Nitrosomonas sp.) oxidize ammonium (NH₄⁺) to nitrite (NO₂⁻) and NO₂⁻ is further converted to nitrate (NO₃⁻) by secondary nitrifiers (Nitrobacter sp.) (Giguere et al. 2018; Chen et al. 2019b). Oxidation of ammonia leads to nitrification-driven N₂O emissions (Hu et al. 2015). Denitrification occurs under anaerobic conditions in which NO₃⁻ is stepwise transformed into N₂O and N₂ by facultative anaerobic bacteria (Xu et al. 2017; Giguere et al. 2018; Zhang et al. 2018a).

Rice (Oryza sativa L.) is cultivated on nearly 155 million ha around the world and feeds approximately 50% of the world's population (Fahad et al. 2019; Bashir et al. 2020). Paddy soils offer a diverse environment of biochemical cycles due to variation in water management practices (Kimura 2000). Nitrification occurs in the rhizosphere zone of paddy soils where nitrogen fertilizers release NH₃ (Wei et al. 2019; Yao et al. 2020). The NO₃⁻-N and NO₂⁻-N are stepwise transformed into N₂O and N₂ through the denitrification process (Ishii et al. 2009). A significant amount of N₂O is released from paddy soils to the atmosphere during the mid-season drainage period (Liu et al. 2019b; Liao et al. 2020). However, N₂O could be further transformed into N₂ during the anaerobic state of soils that creates an anaerobic environment in paddy soils (Abbas et al. 2019). N₂O emission during one growing season of rice paddy is estimated to be between 9 and 35 Gg N/year (Cai 2012). Previously, most studies have focused on N₂O emissions from agricultural soils, wetlands, and grasslands (Brenzinger et al. 2018; Zhang et al. 2019; Ogle et al. 2020; Žurovec et al. 2021). However, few studies have reported the N₂O emission from rice paddy at a national scale.

Recently, process-based models such as denitrification decomposition (DNDC) have been developed to add the feature of estimating N₂O emissions at a regional scale (Gaillard et al. 2018; Tian et al. 2018; Yue et al. 2019). Additionally, these models are able to accurately predict N₂O emissions under different agronomic management practices (Dutta et al. 2018; Goglio et al. 2018; Jiang et al. 2020). The effect of climate change on crop yields can also be assessed using these models (Rafique et al. 2014; He et al. 2019). These models are frequently used for the identification of estimation factors of N₂O emissions (Gaillard et al. 2016; He et al. 2019). Different models such as DayCent, EPIC, CERES-EGC, and SWAT have been developed for these applications (Gaillard et al. 2018; Massara et al. 2018; Giltrap et al. 2020). At first, the DNDC model was explicitly developed for simulation and estimation of N₂O emissions, but now it has been further developed to include other ecosystems (Cui & Wang 2019; Yin et al. 2020; Zhao et al. 2020). The DNDC model can be used to simulate CO_2, CH_4, N₂O, NO, and NH₃ emissions (Shen et al. 2018; Jiang et al. 2020; Pandey et al. 2021). DNDC was previously used to estimate N₂O emissions from agricultural fields (Abdalla et al. 2020b; Jiang et al. 2021; Pandey et al. 2021), animal farms (Brown et al. 2001; Shen et al. 2019), and CH₄ emissions from paddy fields (Li et al. 2002; Wang et al. 2021b). This model connects ecological drivers such as soil properties, vegetation type, climate, and anthropogenic activities to soil and environmental variables. The links between these variables simulate transformation processes of organic carbon and nitrogen, through which CO_2, CH_4, N₂O, NO, and NH₃ are estimated (Camarotto et al. 2018; Zhang et al. 2018b; Li et al. 2019). The two operating modes of the DNDC model are site and regional modes. The site mode can simulate trace gas emissions at a particular site and therefore can be compared with field observations. The regional mode can estimate trace gas emissions from the entire region which is based on statistical uncertainty estimates.

Previously, specific site and treatment-oriented studies on N₂O emission from rice paddy fields have been reported in South Korea (Berger et al. 2013b; Kim et al. 2014a, 2016a; Pramanik et al. 2014). However, the present study is the first to estimate N₂O emission from rice paddy fields on a national scale in the Republic of Korea (South Korea) using the DNDC model. The objective of this study was to estimate N₂O emissions from rice-growing regions of South Korea and identify biogeochemical factors such as NH₄⁺ and NO₃⁻ that are known to affect N₂O emissions. In this paper we will briefly show the preliminary results of a DNDC model of N₂O fluxes from 17 different rice paddy fields of South Korea.

The official name of South Korea is the Republic of Korea (ROK), located in eastern Asia and covers the southern part of the Korean peninsula, which is located between the Sea of Japan (East Sea) in east, the Yellow Sea in west, and the Korea Strait, a sea passage between South Korea and Japan in the south. South Korea shares a land border with North Korea in north. The country also shares maritime borders with China and Japan (Macdonald & Clark 2018). The total area of South Korea is 99,678 km² the country is about the size of Iceland, or slightly smaller than the United States' State of Pennsylvania (You 2009). South Korea is divided into six provinces namely Chungcheongbuk-do, Chungcheongnam-do, Gyeongsangbuk-do, Gyeongsangnam-do, Jeollabuk-do, and Jeollanam-do, one autonomous province Jeju-do, six metropolitan cities namely Busan, Daegu, Daejeon, Gwangju, Incheon, and Ulsan. Seoul is the capital city of the country. South Korea has a population of 50.8 million inhabitants in 2016. The official language is Korean. Forests cover 64% of the total land area in South Korea (Lee et al. 2019). Rice is the most important crop here, accounting for about 90% of the country's total grain production and over 40% of farm income. Barley, soybeans and potatoes are the other major broad acre crops. Fruit, particularly citrus, and vegetables are also widely grown (Neszmelyi 2017).

N₂O emissions were estimated from 17 rice-growing regions across South Korea, as shown in Table 1. To compile data sets on different management practices of agricultural systems such as (1) soil properties, (2) irrigation, and (3) fertilizer application we referred to reported studies on a rice paddy fields in South Korea as shown in Table 2, and supplementary materials Tables S1, and S2, respectively. Reported agronomic practices varied slightly such as date of sowing, fertilizer application, irrigation, and harvesting were different. For missing data on input parameters, we used average values as prevalent in the region. The date of sowing was 12-06-2019 and the date of harvesting was 18-10-2019. Based on data available in Table 2 we selected soil texture silt loam, average bulk density, soil pH, and soil organic carbon as 1.17 g cm⁻³, 6.2, and 0.0201 (kg C·kg⁻¹), respectively. The date of flooding and the date of end-season drainage were selected from Table S1 and was 11-06-2019 and 29-09-2019, respectively. The dates of fertilizer application at different growing stages of the crop such as basal, tillering, and panicle stage were 11-06-2019, 29-06-2019, and 23-07-2019, respectively (Table S2). Climatic characteristics of each region, such as minimum and maximum temperature, precipitation, wind speed, solar radiation, and air humidity, were collected from Korean metrological administration reports 2019 (KMA 2019).

The DNDC-Rice model has three main segments: (1) soil climate, (2) crop growth, and (3) biogeochemistry of soil. The scientific aspects and background information of this model are explained by Fumoto et al. (2008). The DNDC model is a process-based computer simulation model based on carbon and nitrogen biogeochemistry that was originally developed to estimate N₂O emissions in agroecosystems. The model is based on two important components such as physicochemical and biochemical components. The physicochemical component includes (soil climate, crop growth, and decomposition sub-models) that predicts the redox potential, pH, soil temperature, soil moisture, and substrate concentration profiles driven by ecological drivers such as climate, soil, vegetation, and anthropogenic activities. The biochemical component includes nitrification, denitrification, and fermentation sub-models and predicts the gas emissions from plant–soil systems such as carbon dioxide, nitrous oxide, nitric oxide, dinitrogen, methane, and ammonia (Wang et al. 2022).

All the input factors in each component of the DNDC-Rice model are reported in Fumoto et al. (2010). The soil climate simulates the functions of moisture, temperature, and O₂ in the soil layers (0–50 cm depth) and it also simulates the greenhouse gases based on physicochemical properties of soil, daily weather, and agronomic practices. The movement of oxygen in undisturbed soil cores of croplands was explained with Buckingham–Burdine–Campbell diffusion model (Osozawa & Kubota 1987). The crop growth in the DNDC-Rice model is simulated with the computer program MACROS modules (Penning de Vries et al. 1989). In the crop growth section, nitrogen availability, the atmospheric concentration of CO₂, respiration of soil, and allocation of C are considered as main components of the photosynthesis process of a plant (Zhang & Niu 2016). The fluxes of carbon from rice paddy roots to soil are assessed as a part of the C balance. In soil biogeochemistry section, different biogeochemical processes are simulated as a part of soil and environmental properties. The organic carbon decomposition is assessed with first-order reaction kinetics, where the controlling factors are soil water content, temperature, proportion of clay content, concentration of O₂, nitrogen deficiency, and soil tillage. In this section, hydrogen (H₂) and dissolved organic carbon (DOC) production is calculated under anaerobic conditions. Both H₂ and DOC are worked as electron donors for the reduction of oxidants such as Mn⁴⁺, Fe³⁺, and SO₄²⁻ while helping in the production of CH₄. When paddy soil is drained, both oxidants and CH₄ are oxidized by atmospheric O₂ diffusion into paddy soil and results in N₂O emissions.

The simulated results of this study were validated using the experimental records of field studies conducted in South Korea as published previously (Berger et al. 2013a; Kim et al. 2014a). The same procedure of validation of DNDC simulations was adopted by Cai et al. (2003) and Fumoto et al. (2008). Berger et al. (2013a) conducted field experiments and N₂O flux measurements that were done between 11 May 2010 and 23 October 2010 at the FDFM paddy and between 6 May and 15 September 2011 at all three paddies. To measure N₂O exchange at the soil–water/atmosphere interface, closed-chamber measurements in conjunction with a photoacoustic infrared gas analyses (Multigas Monitor 1312, INNOVA, Ball-erup, Denmark) was used every 2 days at each experimental site. One day before the measurement, eight polyvinylchloride (PVC) cylinders (20 cm long and 19.5 cm wide) were installed 6 cm deep in the soil, so that depending on the water level of the rice paddy they poked out of the paddy water at least 2 cm. At each rice paddy, four of them contained rice plants, the other four were installed on spots without rice plants. For the measurement days, the cylinders were connected to chambers with a tubing connection to the gas analyzer, which determined the N₂O concentration of the chamber's headspaces after 0, 8, 16, 24, and 36 min. The reproducibility of one single N₂O concentration measurement was±32 ppb. From a linear increase or decrease of the N₂O concentration in the chambers’ headspaces the N₂O flux was calculated considering the total chamber volume of the gas analyzing system, including the chamber head-space volume, volume of the two 25 m long Teflon tubes and of the CO₂ and H₂O gas traps. Cumulative N₂O emissions were calculated by multiplying the N₂O emission rates of 2 consecutive measurement days with the corresponding time period. These time-weighted N₂O flux means were then summed up over the measurement period.

Kim et al. (2014a) conducted a field experiment in which rice cultivation experiments were carried out at the National Academy of Agricultural Science Research Farm (37°15′N; 126°59′E; 12 m elevation), Rural Development Administration, Suwon, Korea, in 2008. Methane and N₂O emission characteristics were investigated during the rice cropping season using the closed-chamber method (Rolston 1986). In each plot, three transparent acryl chambers (width 62 cm, length 62 cm, and height 112 cm) were placed permanently into the soil after transplanting the rice seedlings. There were four holes in the bottom of each chamber to control the amount of floodwater. The chamber was equipped with a circulating fan to ensure complete gas mixing during the period of sampling. Eight rice plants were covered by each chamber. Gas sampling was carried at 11:00–13:00 to determine the average daily N₂O emission rates during the cropping season. Briefly, gas samples in triplicates were collected once a week using 50-mL air-tight plastic syringes at 0-, 15-, and 30-min intervals after manually closing the chamber. The collected gas samples were transferred into 30-mL air-evacuated glass vials sealed with a butyl rubber septum. Nitrous oxide concentrations were determined using gas chromatograph (Shimadzu, GC-2010) with a stainless-steel column packed with Porapak Q column (Q 80–100 mesh) and equipped with a 63 Nielectron capture detector. The temperatures of the column, injector, and detector were adjusted at 70, 80, and 320 °C, respectively. Helium and hydrogen gases were used as the carrier and burning gases, respectively.

The SPSS Statistics 20 software package was used for analysis. The Pearson correlation test was performed to determine the correlation between simulated and observed N₂O emission.

Rice paddies were cultivated in 17 different regions of South Korea as shown in supplementary material Figure S1. Rice paddies are cultivated on 754,713 hectares that correspond to 87% of the total arable land of South Korea (Table 1). A maximum 161,442 hectares of rice paddies was cultivated in Jeollanam-do and a minimum of 113 hectares in Jeju-do. In each region N₂O emission was observed only at basal, supplementary one, and supplementary two fertilizer applications and after end-season drainage. N₂O emission after the dates of basal N fertilizer application was reported (Chen et al. 2019a). N₂O emission was not observed between all three fertilizer applications. It was also not observed between supplementary fertilizer two and end-season drainage. Oxidation and reduction take place in paddy fields. The application of ammonium N containing fertilizers is nitrified in the oxidized layer, at the water–soil interface, forming nitrate, which moves downwards and denitrified in the reduced layer, producing N₂O. In the presence of a water layer on the soil N₂O is further reduced and escapes to the atmosphere as N₂. N₂O is emitted mainly through the soil surface in the absence of floodwater (Xing et al. 2009; Majumdar 2013; Wu et al. 2022). Since there was continuous flooding between three fertilizer applications and between supplementary fertilizer and end-season drainage N₂O was further reduced and escaped to the atmosphere as N₂, therefore N₂O emission was not observed during these periods of the growing season. Minimum and maximum N₂O emission were observed at basal fertilizer application and after end-season drainage, respectively. N₂O emission varied significantly due to the change in climatic factors and physicochemical properties of each region. The continuous flooding, periodic drainage, seasonal tillage, and other agricultural activities during the rice-growing season make paddy soils in a state of dry–wet alternation (Liao et al. 2020; Zuo et al. 2022). Consequently, different soil layers have different water content and aeration statuses, leading to the changes in chemical composition such as soil redox, pH, state of nitrogen, and other properties in time and space (Wang et al. 2021a; Zuo et al. 2022). In addition, environmental changes such as the changes in soil moisture, temperature, can affect the composition of soil microbial communities, thus, in turn, affecting the nitrification and denitrification of soil that may further lead to either increases or decreases in N₂O emission (Wang et al. 2017; Qin et al. 2020).

Using the DNDC model, we simulated N₂O emissions in rice paddies of all rice paddy growing regions of South Korea. Reported studies in South Korea and simulated N₂O emissions in this study were strongly correlated R² =0.88–0.90. The difference between reported and simulated N₂O emissions was 5–6%. N₂O emissions in all rice paddy growing regions of South Korea were mainly due to NH₄⁺ and NH₃- during flooding (wet period) and after end-season drainage (dry period), respectively. Minimum and maximum N₂O emissions were observed at basal fertilizer application and during end-season drainage, respectively. There was no significant difference between the reported average N₂O emissions in rice paddies worldwide, in South Korea, and that assessed in this study. This comparison of N₂O emissions reflected the reliability of N₂O estimation and our simulated results are within the range of reported studies.

The authors are thankful to the Higher Education Commission (HEC) Pakistan for providing a scholarship for Ph.D. studies at Hanyang University, Seoul, South Korea.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.