Effect of Surface Methane Controls on Ozone Concentration and Rice Yield in Asia

Abstract

Surface methane (CH₄) is a significant precursor of tropospheric ozone (O₃), a greenhouse gas that detrimentally impacts crops by suppressing their physiological processes, such as photosynthesis. This relationship implies that CH₄ emissions can indirectly harm crops by increasing troposphere O₃ concentrations. While this topic is important, few studies have specifically examined the combined effects of CH₄ and CH₄-induced O₃ on rice yield and production. Utilizing the GEOS-Chem model, we assessed the potential reduction in rice yield and production in Asia against a 50% reduction in anthropogenic CH₄ emissions relative to the 2010 base year. Based on O₃ exposure metrics, the results revealed an average relative yield loss of 9.5% and a rice production loss of 45,121 kilotons (Kt) based on AOT40. Regions such as the India-Gangetic Plain and the Yellow River basin were particularly affected. This study determined that substantial reductions in CH₄ concentrations can prevent significant rice production losses. Specifically, curbing CH₄ emissions in the Beijing-Tianjin-Hebei region could significantly diminish the detrimental effects of O₃ on rice yields in China, Korea, and Japan. In summary, decreasing CH₄ emissions is a viable strategy to mitigate O₃-induced reductions in rice yield and production in Asia.

1. Introduction

Fossil fuel combustion is a major source of carbon emissions and causes climate change [1,2,3]. Methane (CH₄) specifically serves as a potent yet short-lived climate pollutant, playing a marked role in the generation of ground-level ozone (O₃), impacting human and ecosystem health [4]. CH₄ concentrations have increased by over 150% since the pre-industrial era, with anthropogenic activities now accounting for about 50% more emissions than natural sources. In the last decade, CH₄ concentrations have increased rapidly, with the highest growth rate recorded in 2020 [5]. This increase is attributed to emissions from agriculture, fossil fuel production, solid waste in landfills, and wastewater management [6]. Anthropogenic CH₄ emissions are expected to continue to increase, potentially reaching approximately 380 million tons per year by 2030, an 8% increase from 2020 levels [7]. These emissions vary regionally depending on fossil fuel sources, waste management systems, and agricultural practices, making CH₄ mitigation crucial for addressing climate change, human health, ecosystem, and agriculture-related concerns [8].

O₃ is produced when sunlight interacts with NO_x, CO, NMVOC, and CH₄ emissions, notably reducing crop yields and quality. Liu and Desai [9] reported that relative crop yield losses due to air pollution from O₃ and aerosols have ranged from 20 to 30% in the United States over the past four decades. Additionally, O₃ exposure results in global yield losses of 7.1, 12.4, 6.1, and 4.4% for wheat, soybean, maize, and rice, respectively [10,11,12,13]. For every 1 million tons of CH₄ reduction, there can be a yield loss prevention for 55,000, 17,000, 42,000, and 31,000 tons of wheat, soybeans, maize, and rice, respectively [13]. Shindell et al. [11] demonstrated that a 50% reduction in anthropogenic CH₄ emissions would lead to a 134 million ton decrease in CH₄ emissions, preventing 4.2 million tons of rice yield losses. Major rice producers such as India, China, Bangladesh, and Vietnam would see reduced losses in rice production due to these decreased emissions [5]. It is important to assess the effect of reducing anthropogenic CH₄ emissions on crop yields, and atmospheric chemical models are crucial for estimating CH₄ emissions and their impact on surface O₃ concentrations. However, these studies often concentrate on individual countries or specific sectors, thereby lacking a comprehensive analysis that combines the effects of CH₄ and CH₄-induced O₃ on rice productivity on a regional scale, especially in Asia.

Based on these insights, this study focuses on examining the detailed relationship between CH₄-induced O₃ and rice productivity in Asia. While previous studies have mainly focused on modeling at the country level, only a few have attempted to analyze the response of O₃ to CH₄ emissions and rice yield in Asia using atmospheric chemistry models to simulate sub-grid scale data. Therefore, there is a notable gap in understanding their combined impact on rice productivity in Asia. This study is unique as it evaluates the relationship between O₃ and rice yield and production under a one-half anthropogenic CH₄ emission scenario. The specific objectives of this study are to (1) investigate the applicability of the GEOS-Chem model in predicting O₃ responses to anthropogenic CH₄ emission; (2) analyze O₃ exposure metrics such as accumulated O₃ exposure over a threshold of 40 ppb (AOT40) and mean 7 h O₃ mixing ratio (M7); (3) quantitatively evaluate rice yield and production loss in Asia due to CH₄-induced O₃ exposure. The framework of this study involves data collection, model simulation, and analysis. Initially, I will collect and analyze existing data on CH₄ emissions and O₃ concentrations. Subsequently, I will employ the GEOS-Chem model to simulate the effects of reduced anthropogenic CH₄ emissions on O₃ levels and, consequently, on rice yield and production in Asia. Finally, based on these findings, the policy implications of this study will be discussed.

2. Materials and Methods

2.1. Model Description

The GEOS-Chem model (version 13.3.4) was used following the approach outlined by Bey et al. [14] to comprehensively analyze the impact on rice yield and production resulting from a 50% reduction in anthropogenic CH₄ emissions. The baseline for comparison was set in 2010 (

Table 1. Simulation cases of methane emissions were used in this study.

Simulation Name	CH4 Emissions or Mixing Ratio (ppbv)	Reference
BASE	1808	World Meteorological Organization [15]
CASE1	50% anthropogenic CH4 reduction	Fiore et al. [16]

). The GEOS-Chem model is an advanced global three-dimensional chemical transport model designed to simulate the composition of the Earth’s atmosphere. It operates using meteorological data sourced from the Goddard Earth Observing System (GEOS), managed by NASA’s Global Modeling and Assimilation Office (GMAO). This model enables a nuanced understanding of how changes in CH₄ emissions can affect atmospheric O₃ levels, subsequently influencing rice yield and production across various regions.

To conduct the simulations, we organized the GEOS-Chem model into two domains for nested-grid simulation. The first domain covered the entire globe with a 4° × 5° grid resolution and 72 vertical layers, spanning from 1 January 1990 to 31 December 2010 (UTC). The year 2010 was selected as the reference year due to the availability of reliable emission and rice yield data up to that year. The second domain focused on East, South, and Southeast Asia, with a higher resolution of 0.5° × 0.625° and 72 vertical levels, covering the period from 1 January 2009 to 31 December 2010 (UTC). The nested simulations used data from the global run to establish lateral boundary conditions. The time step was 300 s for transport and convection and 600 s for chemicals and emissions. The spin-up period was 20 years for the outer domain and 1 year for the interdomain.

2.2. Emission and Meteorological Data

For emissions data, we utilized the Community Emissions Data System (CEDS v2021-06) for global monthly mean anthropogenic emissions, including CO, CO₂, NO_x, SO₂, NH₃, NMVOC, organic/black carbon, and emissions from various sectors such as agriculture, energy, industry, and transportation. The CH₄ mixing ratio data was obtained from the World Meteorological Organization [15]. Other emission sources such as burning, dust, sea salt, lightning NO_x, and soil NO_x were also taken into account from relevant studies [17,18,19,20,21]. To calculate biogenic emissions, the Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN 2.1) was utilized. MEGAN 2.1 estimates the release of isoprene, monoterpenes, and various other trace gases and aerosols from ecosystems into the atmosphere [22]. For a detailed description of the GEOS-Chem emissions, please refer to Bey et al. [14]. The framework for CH₄ emissions, which includes the percentage of anthropogenic emissions among CH₄ emissions and the relationship between CH₄ emissions and concentrations with the OH feedback coefficients, followed the procedure described in Prather et al. [23], Prather [24], and Fiore et al. [16].

Initial meteorological and boundary conditions were obtained from the modern-era Retrospective analysis for Research and Applications, version 2 (MERRA-2) dataset [25]. This is a global atmospheric reanalysis produced by the GMAO with a horizontal resolution of 0.5° × 0.625° and 72 hybrid sigma/pressure levels. Figure 1 shows a schematic of the simulation process across Asia.

2.3. Observation Dataset

The 2010 rice production data utilized in this study were sourced from the Global Agro-Ecological Zones (GAEZ) inventory [26]. While the Food and Agricultural Organization’s FAOSTAT database provides national-level agricultural data, it lacks the level of detail needed to capture spatial variations. To address this limitation, we utilized GAEZ data, which downscales national production statistics to a grid scale by integrating various geospatial data, including remote sensing, soil, climate, and population density. The GAEZ data, available in a 5 arc-minute raster grid format, were adjusted to match the GEOS-Chem model grid. Additionally, the Crop Calendar Dataset [27] supplied gridded rice planting and harvesting dates, which were crucial for computing the AOT40 and M7 metrics during the growth period.

To validate the accuracy of the model outputs across Asia, observed surface O₃ concentrations were obtained from the MACC global reanalysis of assimilated gridded O₃ data at the surface [28] for comparing the modeled surface O₃.

2.4. Rice Yield and Production Losses Based on Ozone Exposure Metrics

O₃ exposure metrics for the 2010 growing season were computed as rice’s sensitivity to O₃ significantly decreases after panicle formation [29]. The exposure-response function varies according to the region, and statistical methods and definitions of the growing season differ. Consequently, assessing the suitability of using exposure metrics in regions other than their original context is complex, given the considerable uncertainties in estimating rice yield losses via these metrics [30,31]. To address these complexities, we employed both AOT40 and M7 metrics to predict rice yield losses based on exposure metrics. While AOT40 is primarily utilized in Europe to assess plant risk from O₃ exposure and estimate crop yield losses across various regions [32,33], the M7 metric measures the daily mean surface O₃ concentration during a specific 7 h period throughout the growing season of the plant. Equations (1) and (2) were utilized to calculate the AOT40 and M7 metrics, respectively.

$$\mathrm{AOT}40={\displaystyle {{\displaystyle \sum }}_{i=1}^n}\left({\left[{\mathrm{O}}_3\right]}_i-0.04\right),\mathrm{for} {\mathrm{O}}_3\ge 0.04 \mathrm{ppmv} \mathrm{from} 8:00 \mathrm{to} 19:59 \left(\mathrm{LST}\right),$$

(1)

$$\mathrm{M}7=\frac{1}{n}{\displaystyle {{\displaystyle \sum }}_{t=09\text{:}00}^{t=15\text{:}59}}{{[\mathrm{O}}_3]}_t,$$

(2)

where ${{[\mathrm{O}}_3]}_i$ denotes the hourly mean surface O₃ concentration in parts per million by volume (ppmv), n is the total number of hours in the growing season, and ${{[\mathrm{O}}_3]}_t$ signifies hourly mean surface O₃ concentration in parts per billion by volume (ppbv) from 9:00 to 15:59 LST.

The relative yield (RY) of rice was estimated and subtracted from unity to calculate O₃-induced relative yield loss (RYL), expressed as follows:

RYL = 1.0 − RY,

(3)

where RY is the relative rice yield with O₃ damage, and RYL is the theoretical reduction in rice yield that would have occurred with O₃-induced damage. RY was estimated based on the empirical relationships derived for AOT40 [34] and M7 [32].

RY = 1.0 − (0.0045 × AOT40),

(4)

$$\mathrm{RY}=\exp \left[\right(-(\mathrm{M}7/{202)}^{2.47}]/\exp [-(25/202{)}^{2.47}].$$

(5)

The rice production loss (RPL) was computed using Equation (6) for each grid cell i in the rice cultivated area using RYL and the actual rice production for 2010 obtained from GAEZ:

$${\mathrm{RPL}}_{\mathrm{i}}{=\mathrm{CP}}_i{\times \mathrm{RYL}}_i{/(1-\mathrm{RYL}}_i)$$

(6)

where CP symbolizes the actual rice production.

3. Results and Discussion

3.1. Reproduction Accuracy of Simulated Surface Ozone Concentration

Figure S1 shows the relative error of the monthly mean surface O₃ for BASE simulation with respect to the MACC global reanalysis of assimilated O₃ for 2010 [28]. The relative error was within ±15% in almost the entire simulated domain. In many cities and industrial regions across Asia, emissions of O₃ precursors, specifically NO_x and NMVOC, are prevalent, often promoting O₃ formation with elevated O₃ concentrations reported in urban areas of China and India. Moreover, O₃ concentrations exhibit seasonal variations in many parts of Asia, where strong solar radiation enhances O₃ formation during summer. Here, monthly simulated O₃ values on land were overestimated by up to 15% from July to September. Conversely, O₃ values at sea were underestimated by 0–10% throughout the year. Similar trends were found in the model inter-comparison study [35]. These results may stem from (1) the TP ozone valley [36,37], (2) the depiction of the dispersion of southwesterly clean marine air masses [38], and (3) the considerable diversity in O₃ photochemical production [35,39]. Major challenges remain with regard to reproducing the surface O₃ over Asia using the GEOS-Chem model. However, the relative error of surface O₃ is generally acceptable. Here, I evaluated the effect of O₃ on CH₄ emission controls for crops based on the simulation results.

3.2. Surface Methane and Ozone Distributions

Figure 2 shows the average surface CH₄ and O₃ mixing ratio of the rice cultivated area for major rice producer countries monthly for the BASE simulation. Figures S2–S4 show the simulated spatial distribution of CH₄, O₃, and OH mixing ratios, respectively, at the ground level for each month for the BASE 2010 simulation. The CH₄ mixing ratio in China, Korea, and Japan was relatively high compared with those of other countries (Figure 2a), whose CH₄ mixing ratio was relatively high in winter compared with summer (Figure 2a), attributed to OH radicals (Figure S4) produced using ultraviolet light reduction during summer. The surface CH₄ mixing ratio in summer was particularly high in Beijing-Tianjin-Hebei (BTH), North China Plain (Figure S2), reflecting anthropogenic CH₄ emissions from coal, natural gas, and landfills [40]. The relatively high CH₄ in summer and baseline CH₄ in China, Korea, and Japan is mainly attributed to air transportation of the relatively high CH₄ mixing ratio from BTH to Korea and Japan. Except for China, Korea, Japan, and Pakistan, the O₃ mixing ratio peaks from spring to winter (Figure 2b and Figure S3); however, the summer O₃ baseline in these four countries is high compared with that of other countries. The relatively high winter-spring O₃ concentrations in all countries except China, Korea, Japan, and Pakistan are mainly attributed to O₃ photolysis followed by the reaction of excited oxygen atoms (O(¹D)) with a relatively high water vapor content in summer and the reaction of OH radicals (Figure 2c) leading to the loss of O₃. Moreover, the inflow of oceanic clean-up associated with the Asian summer monsoon would reduce O₃ concentration during summer in a relatively low latitude zone. The O₃ mixing ratio peak during summer in China, Japan, and Korea was mainly attributed to (1) the inflow of O₃ from the south to the north by the Asian summer monsoon (Figure S3) and (2) the higher CH₄ emission at BTH (Figure S2). Furthermore, one of the reasons for the high summer O₃ concentration in Pakistan is the influence of the Karakoram range, Hindu Kush mountains, and Himalayas on the summer southwest monsoon, resulting in the retention of high O₃ concentrations (Figure S3). These O₃ trends are consistent with those reported by Liu et al. [41].

3.3. Distributions of Surface Accumulated Ozone Exposure

The Indo-Gangetic Plain and North China Plain have relatively high summer surface O₃ concentrations [42,43] owing to the comparatively large production of rice across India and China, respectively (

Table 2. Relative yield losses (RYL; %) and rice production losses (RPL; Kt) based on AOT40 (ppmh) and M7 (ppbv) metrics for the year 2010, according to country.

Country	BASE			CASE1			BASE			CASE1
	AOT40	RYL	RPL	AOT40	RYL	RPL	M7	RYL	RPL	M7	RYL	RPL
Japan	9.1	3.8	236.1	6.8	2.8	172.8	45.5	2.0	121.5	43.2	1.6	101.2
Republic of Korea	13.7	5.6	282.5	10.9	4.4	219.8	50.3	2.6	131.8	47.6	2.2	110.5
North Korea	15.5	6.4	136.7	12.4	5.1	108	48.2	2.3	49.1	46.0	2.0	41.8
China	26.5	10.8	22,202.0	22.3	9.1	18,296.1	55.8	3.7	7047.2	52.9	3.2	5971.4
Philippines	0.2	0.1	8.0	0.1	0.1	5.9	21.8	0.1	7.2	20.7	0.0	4.4
Vietnam	11.2	4.7	1138.5	9.7	4.1	949.1	41.8	1.6	416.1	39.8	1.4	348.0
Cambodia	5.6	2.3	74.1	4.2	1.7	54	34.6	0.7	23.6	32.5	0.5	16.8
Laos	19.5	8.7	94.5	17	7.5	81.9	47.1	2.4	22.9	45.1	2.1	20.0
Thailand	3.3	1.4	319.6	2.5	1.1	237.3	31.5	0.6	118.8	29.5	0.4	87.7
Myanmar	1.0	0.4	21.9	1	0.4	26.7	24.0	0.1	5.0	22.1	0.1	2.5
Malaysia	6.3	3.3	38.8	5.4	2.8	32.4	32.4	0.9	12.0	31.0	0.8	10.3
Indonesia	1.6	0.7	541.5	1.3	0.6	428.8	25.6	0.3	215.4	24.5	0.3	177.9
Bangladesh	8.0	3.3	1183.3	5.9	2.4	804.5	38.2	1.2	490.6	34.9	0.8	338.8
Nepal	14.9	6.3	363.6	10.6	4.4	256.2	47.7	2.4	115.5	44.0	1.9	90.9
Bhutan	41.5	18.6	10.3	35.1	15.8	8.4	61.4	4.8	2.2	58.1	4.1	1.9
India	32.8	13.8	17,322.4	26.8	11.2	13,645.8	53.9	3.5	3861.4	50.5	2.9	3161.7
Pakistan	32.3	13.6	1116.8	25.2	10.6	815.3	58.0	4.3	324.9	53.9	3.5	256.5
Brunei	0.1	0.0	0.0	0.1	0.0	0.0	25.3	0.0	0.0	24.5	0.0	0.0
Taiwan	10.6	4.7	26.8	7.9	3.5	20.2	46.9	2.3	12.8	43.8	1.8	10.3
Sri Lanka	0.6	0.3	2.9	0.3	0.2	1.4	31.4	0.4	8.0	29.7	0.3	5.4
Asia	22.3	9.5	45,120.5	18.4	7.8	36,164.6	48.5	2.9	12,985.9	45.8	2.4	10,757.9

). These regions also experience the highest O₃ pollution levels (Figure S3). Elevated O₃ precursor levels in these areas—influenced by wind convection, clear skies, high pressure, and pollutant circulation—contribute to increased O₃ formation and accumulation. Figure 3 shows the spatial distribution of AOT40 and M7 in BASE simulation. Figure S5 shows the spatial distribution of AOT40 and M7 for CASE 1. Averaged AOT40 (M7) in rice cultivated areas under the BASE scenario reached 22.3 ppmh (48.5 ppbv), and the reduction rate of AOT40 (M7) for the simulation with CASE 1 was 17% (6%) (Table 2). Both metrics were relatively high between 25–40° latitude for all scenarios, with a hotspot identified in the Indo-Gangetic Plain and Yellow River basin. Conversely, the southern parts of India and Southeast Asia showed relatively low AOT40 and M7 metrics. Deb Roy et al. [44] indicated that the AOT40 values in 2003 were higher along the Himalayas. Feng et al. [45] and Macro et al. [46] found that AOT40 levels were relatively high north of the Yellow River basin compared with central and south China. However, full comparisons are complicated by the differences in the simulation models and years; however, the results obtained here exhibit trends similar to those observed in previous studies. One of the factors contributing to the relatively low AOT40 at lower latitudes is the inflow of air with low O₃ concentrations caused by the southwest summer monsoon.

Solar radiation through clear skies, NO₂ photolysis, and photo-oxidation of NMVOC (promotion of NO₂) exert a notably high positive correlation between CH₄ and O₃ in summer (not shown). Therefore, reducing the CH₄ emissions during the summer effectively reduces the impact of O₃ on crops. Although high O₃ concentrations in the India-Gangetic Plain are attributed to meteorological and geographical conditions, in the BTH region, CH₄ emissions from major industries such as coal, agriculture, and petroleum are likely the major contributors to these concentrations. CH₄ emission reductions in BTH would reduce AOT40 and M7 in the 30–40° latitude band.

3.4. Relative Rice Yields and Production Losses

Figure 4 shows the O₃-induced RYL based on AOT40 and M7 for BASE and CASE 1 simulations. The spatial distribution of RYL based on AOT40 closely aligns with the results reported by Sharma et al. [47] and Cao et al. [48]. According to our results, all rice cultivation areas experienced some degree of damage and yield reduction for both metrics. Based on the country, Aunan et al. [49] showed that RYL based on M7 in China ranged from 1.1 to 1.5%. Wang and Mauzerall [32] indicated that RYLs based on M7 in China, Japan, and South Korea are 3–5, 4, and 2%, respectively. Van Dingenen et al. [33] indicated that RYLs based on AOT40 (M7) in China and India are 3.9% (3.1%) and 8.3% (5.7%), respectively. Wang et al. (2012) indicated that the RYL in China based on M7 is 5%. Sinha et al. [50] showed that RYLs based on AOT40 in China ranged from 12 to 14%. Danh et al. [51] showed that RYLs based on AOT40 (M7) in Vietnam ranged from 0.4 to 5.9% (0.02 to 0.06%). Lal et al. [52] showed that the RYL based on AOT40 (M7) in India was 6.7% (0.3%). Lin et al. [53] indicated that RYLs based on AOT40 in China ranged from 11 to 17%. Sharma et al. [47] showed that RYLs based on AOT40 in India were less than 6%. Zhao et al. [54] indicated that RYLs based on AOT40 in China ranged from 3.9 to 7.3%. For this study, the aggregated averaged RYLs based on AOT40 (M7) for BASE simulation in China and India were 10.8% (3.7%) and 13.8% (3.5%), respectively (Table 2). The RYL based on AOT40 in China obtained from this study was higher than those reported by Van Dingenen et al. [33] and Zhao et al. [54] but consistent with those reported by Lin et al. [53]. Considering the M7 index, the outcomes demonstrated broad consistency across multiple studies, except for the findings of Aunan et al. [49]. In contrast, when assessing RYLs based on the AOT40 metric in India, they notably exceeded the results reported by Van Dingenen et al. [33], Lal et al. [52], and Sharma et al. [47] but aligned with those reported by Sinha et al. [50]. The RYL based on M7 in India exceeded that reported by Lal et al. [52] but was smaller than that reported by Van Dingenen et al. [33]. The variation in RYLs was attributed to the differences in O₃ distribution, simulation models, growing period, and years. The RYL for rice was largest in Bhutan (18.6% on AOT40; 4.8% on M7), followed by India (13.8% on AOT40; 3.5% on M7), Pakistan (13.6% on AOT40; 4.3% on M7), China (10.8% on AOT40; 3.7% on M7), and Laos (8.7% on AOT40; 2.4% on M7). Throughout Asia, the average RYL was 9.5% on the AOT40 and 2.9% on the M7.

An RYL gradient, moving from north to south, was evident for both East and South Asia, based on estimates from both AOT40 and M7 metrics. (Figure 4). Based on regions, the O₃-induced RYL based on AOT40 (M7) for BASE was largest (RYL > 20% (10%)) in the North China Plain—a highly industrialized and populated area—and a southern part of the Himalayas, which is located at the southern monsoon winds, is blocked by a mountain range, for all simulation scenarios. The rest of the rice cultivated area showed that RYL on AOT40 (M7) ranged from 0 to 20% (0 to 10%). It is noteworthy that areas with higher RYL estimated using the M7 metric closely mirrored those estimated using the AOT40 metrics, but RYL on M7 was 20% lower in all cultivated areas compared with AOT40. The falling rate of RYL in Asia for CASE 1 compared with BASE simulation was 1.7% (0.5%) based on AOT40 (M7), respectively (Figure 4 and Table 2). The RYL value exceeded 5% on AOT40, considered a critical level [45], and was 68.2% for BASE and 59.6% for CASE 1 of the total rice cultivated area. Similarly, the RYL values exceeded 5% on M7 and were 17.2% for BASE and 9.3% for CASE 1.

Considering production, the largest RPL area was the India-Gangetic Plain and the Yangtze River basin, which were slightly south of the larger RYL area owing to the overall higher production in the south (Figure 5). The largest aggregated RPL for BASE simulation was observed in China (22,202 Kt on AOT40; 7047 Kt on M7), followed by India (17,322 Kt on AOT40; 3861 Kt on M7), Bangladesh (1183 Kt on AOT40; 491 Kt on M7), Vietnam (1139 Kt on AOT40; 416 Kt on M7), and Pakistan (1117 Kt on AOT40; 325 Kt on M7) (Table 2). The RPL reduction in Asia for CASE 1 compared with BASE simulation was 8956 Kt on AOT40 and 2228 Kt on M7, respectively (Figure 5 and Table 2). This indicates that a 50% reduction in anthropogenic CH₄ would reduce losses from the total production in Asia by as much as 1.0% based on AOT40 and 0.3% based on M7.

CH₄-induced O₃ affects rice yield and production negatively. Therefore, reducing CH₄ emissions not only mitigates greenhouse warming but also reduces yield losses due to O₃. The highest projected RYL in countries like Bhutan, India, Pakistan, China, and Laos, based on the AOT40 metric, is likely due to anticipated O₃ exposure exceeding 40 ppmh during the growth phase. However, it is important to note that the accuracy of this value is uncertain due to the lack of high-resolution O₃ studies in Asia for 2010 as a basis for comparison. The study reveals a discrepancy between the RYLs calculated using the AOT40 and M7 metrics. Rice appears to be more O₃-resistant according to the M7 metric but sensitive according to the AOT40 metric. This could imply that rice is more vulnerable to short-term, high O₃ levels than to prolonged, moderate levels, which is consistent with the findings of Hollaway et al. [55]. Ground level O₃ is harmful to rice, and reduced CH₄ concentrations increase rice yields.

3.5. Uncertainty from Impacts of Methane Emissions

O₃ is generated using the interaction of sunlight with emissions of NO_x, CO, NMVOC, and CH₄. Furthermore, a certain percentage of tropospheric O₃ is transported from the stratosphere. However, there is approximately a 50% uncertainty in the change in O₃ owing to CH₄ over 40 years calculated using multiple models that allow for the estimation of interactions with NO_x and NMVOC degradation chemistry and the resulting radical levels [56]. Notably, 55% of the increase in the O₃ budget since the pre-industrial era is ascribed to NOx, with 25% linked to CH₄ and 19% to CO and NMVOC [57]. Therefore, CH₄ decomposition chemistry and the resulting O₃ formation are elucidated using laboratory, field, and modeling knowledge. Though model studies indicate contributions from CH₄ to O₃ trends, determining whether model performance limits the ability to predict O₃ trends is difficult [58,59]. Therefore, the extent to which O₃ trends can be attributed to surface CH₄ concentration and estimations of the effect of CH₄-induced O₃ on crop damage are uncertain. The findings of this study provide valuable insights for policymakers, aiding the formulation and implementation of emission control measures targeting O₃ precursors to enhance the resilience of future crop yields. While external factors like weather and soil conditions directly impact rice growth, there is a noticeable research gap concerning the synergistic effects of CH₄ and O₃ levels with meteorological factors, including climate change, on local-scale agricultural production. Therefore, aspects such as long-term surface and troposphere O₃ and CH₄ observations and model development, including parameterization and process improvement, should be studied further.

4. Conclusions

In this study, the GEOS-Chem model was applied to study rice impacts from CH₄-induced O₃, as well as the effect of reducing anthropogenic CH₄ emissions on rice yield and production across Asia. The average O₃-induced RYL for rice in Asia was 9.5% on AOT40 and 2.9% on M7. The O₃-induced damage to rice based on AOT40 and M7 were approximately 45,121 Kt and 12,986 Kt, respectively, on a production basis in Asia. Regionally, the Hindustan Plains to the Deccan Plateau and the Yellow River Basin exhibited relatively high RYLs. The Hindustan Plains and Yangtze River basin exhibited a relatively high RPL. Moreover, simulation analysis shows that reducing anthropogenic CH₄ emissions mitigates O₃ damage to rice production. In Asia, the reduced rice production losses with a 50% reduction in anthropogenic CH₄ emission were 8956 Kt for AOT40 and 2228 Kt for M7. Moreover, CH₄ control in the BTH in the North China Plain would reduce the effect of O₃ pollution on rice yields in high-latitude zones and emphasize the importance of reducing CH₄ emissions. The study findings are expected to contribute to the identification of a more effective and environmentally friendly strategy for improving rice production against the backdrop of CH₄-induced O₃. In addition, this study highlights the pressing necessity for policy initiatives to reduce anthropogenic CH₄ emissions, particularly within sectors such as agriculture, fossil fuel production, and waste management. Implementing such measures can effectively mitigate the adverse effects of climate change and substantially curtail crop yield losses, thereby ensuring food security in major rice-producing nations across Asia.

Future research should concentrate on refining O₃ response models pertaining to CH₄ emissions and improving the accuracy of yield loss predictions. The development of exposure-response relationships with reduced uncertainty will prove pivotal in quantifying the benefits associated with CH₄ emission reductions on rice yields and production.