Next Article in Journal
In Situ Monitoring and Bioluminescence Kinetics of Pseudomonas fluorescens M3A Bioluminescent Reporter with Bacteriophage ΦS1
Previous Article in Journal
A Framework for the Datasets of CRDS CO2 and CH4 Stable Carbon Isotope Measurements in the Atmosphere
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Data
Volume 10
Issue 9
10.3390/data10090151
data-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Validation of Anthropogenic Emission Inventories in Japan: A WRF-Chem Comparison of PM2.5, SO2, NOx and CO Against Observations

by
Kenichi Tatsumi
1,* and
Nguyen Thi Hong Diep
2
1
Graduate School of Data Science, Nagoya City University, Nagoya 467-8501, Japan
2
College of Environment and Natural Resources, Can Tho University, Can Tho City 94115, Vietnam
*
Author to whom correspondence should be addressed.
Data 2025, 10(9), 151; https://doi.org/10.3390/data10090151
Submission received: 23 May 2025 / Revised: 29 August 2025 / Accepted: 18 September 2025 / Published: 22 September 2025
Browse Figures
Versions Notes

Abstract

Reliable, high-resolution emission inventories are essential for accurately simulating air quality and for designing evidence-based mitigation policies. Yet their performance over Japan—where transboundary inflow, strict fuel regulations, and complex source mixes coexist—remains poorly quantified. This study therefore benchmarks four widely used anthropogenic inventories—REAS v3.2.1, CAMS-GLOB-ANT v6.2, ECLIPSE v6b, and HTAP v3—by coupling each to WRF-Chem (10 km grid) and comparing simulated surface PM2.5, SO2, CO, and NOx with observations from >900 stations across eight Japanese regions for the years 2010 and 2015. All simulations shared identical meteorology, chemistry, and natural-source inputs (MEGAN 2.1 biogenic VOCs; FINN v1.5 biomass burning) so that differences in model output isolate the influence of anthropogenic emissions. HTAP delivered the most balanced SO2 and CO fields (regional mean biases mostly within ±25%), whereas ECLIPSE reproduced NOx spatial gradients best, albeit with a negative overall bias. REAS captured industrial SO2 reliably but over-estimated PM2.5 and NOx in western conurbations while under-estimating them in rural prefectures. CAMS-GLOB-ANT showed systematic biases—under-estimating PM2.5 and CO yet markedly over-estimating SO2—highlighting the need for Japan-specific sulfur-fuel adjustments. For several pollutant–region combinations, absolute errors exceeded 100%, confirming that emissions uncertainty, not model physics, dominates regional air quality error even under identical dynamical and chemical settings. These findings underscore the importance of inventory-specific and pollutant-specific selection—or better, multi-inventory ensemble approaches—when assessing Japanese air quality and formulating policy. Routine assimilation of ground and satellite data, together with inverse modeling, is recommended to narrow residual biases and improve future inventories.

1. Introduction

Air pollution is widely recognized as one of the most pressing global environmental issues due to its profound impacts on public health, ecosystems, agriculture, and climate [1,2,3]. Among the regulated pollutants, fine particulate matter (PM2.5), sulfur dioxide (SO2), carbon monoxide (CO), and nitrogen oxides (NOx) are especially critical because of their well-documented effects on air quality and human health [1,4]. In Japan—as in many other countries—these pollutants are rigorously monitored and analyzed to guide mitigation strategies and policy development [5]. Accurate assessments of their spatial and temporal dynamics, however, depend on both reliable emissions data and robust atmospheric chemistry modeling frameworks [6,7].
Emission inventories, which compile source-specific estimates of emissions, are indispensable for air quality modeling and environmental policy planning [8,9]. Regional and global inventories alike are routinely employed to apportion anthropogenic contributions and to drive chemical-transport models [10]. However, inventories often diverge substantially due to differences in methodologies, assumptions, and spatial or temporal resolutions [11]. Such discrepancies affect model simulations, making systematic validation against observations essential [12]. For example, studies that compare PM2.5 model results derived from alternative inventories with ground observations reveal large inter-inventory spreads frequently linked to differences in the treatment of secondary aerosol formation and natural sources [13,14]. Comparable challenges have been reported for NOx and SO2, where contrasting point-source characterizations and emission-factor choices lead to divergent outcomes [11].
In Japan, previous research has tended to focus on single pollutants or sectors. Traffic-related NOx emissions, for instance, remain difficult to quantify accurately in densely populated cities [15]; SO2 estimates are confounded by uncertainties in industrial data and by transboundary inflow from East Asia [16]. These examples underscore the complexity of emissions modeling and the need for comprehensive validation across pollutants and inventories. To date, however, few studies have systematically compared major inventories using nationwide observations across a wide spectrum of pollutants. The present study addresses this gap by evaluating four major global emission inventories—the Regional Emission inventory in Asia (REAS v3.2.1) [9], the Copernicus Atmosphere Monitoring Service Global Anthropogenic Emissions (CAMS-GLOB-ANT v6.2) [7], Evaluating the Climate and Air Quality Impacts of Short-Lived Pollutants (ECLIPSE v6b) [17], and the Hemispheric Transport of Air Pollution (HTAP v3) [11]. We focus on their ability to reproduce Japanese observations of PM2.5, NOx, SO2, and CO for the benchmark years 2010 and 2015.
This study addresses the following research questions:
  • Representation accuracy: How well do REAS, CAMS-GLOB-ANT, ECLIPSE, and HTAP replicate observed PM2.5, NOx, SO2 and CO concentrations in Japan?
  • Regional performance: How does the agreement between each inventory and observations vary across Japan’s major geographic regions?
By systematically benchmarking model outputs against observations, we identify the strengths and weaknesses of each inventory and offer recommendations for improvements.

2. Materials and Methods

2.1. WRF-Chem Description

In this study, the coupled atmospheric-chemistry model WRF-Chem (v 4.6) [18,19] was applied to simulate air-pollutant concentrations—PM2.5, SO2, CO and NOx—over Japan. The model was configured with a single domain covering the entire archipelago, using a horizontal grid of 10 km × 10 km and 35 vertical levels extending from the surface up to 50 hPa. Land-use and topographic data were derived from the MODIS 21-category dataset and GTOPO30, respectively. Meteorological initial and lateral boundary conditions were taken from the NCEP Final Analysis (FNL), which provides 6-hourly fields at 0.25° × 0.25° resolution [20]. Chemical initial and boundary conditions were supplied by the Community Atmosphere Model with Chemistry (CAM-Chem) at 6-hourly intervals and 0.9° × 1.25° resolution [21].
The physical and chemical parameterizations adopted in this study are summarized in Table 1. The gas-phase mechanism was MOZART-4 (Model for Ozone and Related Chemical Tracers) [22], which represents 52 chemical species and 132 photolysis reactions and includes updates for isoprene, monoterpenes, and other volatile organic compounds including ethene, methyl-butanol, and HONO chemistry. Aerosol processes were handled with the MOSAIC (Model for Simulating Aerosol Interactions and Chemistry) module [23], while dust emissions were simulated using the GOCART (Goddard Chemistry Aerosol Radiation and Transport) scheme [24].
For regional analysis, Japan was divided into eight geographic/economic regions (Figure 1): Hokkaido, Tohoku, Kanto, Chubu, Kinki, Chugoku, Shikoku and Kyushu. Hokkaido is characterized by extensive farmland and low population density; emissions arise mainly from livestock operations and wintertime residential heating. Tohoku, likewise predominantly rural, shares similar farming and heating sources. Kanto—the most urbanized region, centered on Tokyo—features emissions from road traffic, commercial/residential fuel use and diverse industries. Chubu is a manufacturing hub where industrial facilities and urban transport dominate. Kinki (Osaka, Kyoto) also shows high industrial and residential-energy emissions. Chugoku includes heavy industry along the Seto Inland Sea and substantial port activity. Shikoku’s rugged terrain and limited urbanization mean emissions stem chiefly from pulp-and-paper mills, agriculture and port operations. Kyushu features both large industrial zones and rural farmland, so emissions include both heavy industry and agriculture.
To evaluate the inventories’ ability to reproduce observed air quality conditions across different temporal baselines, simulations were conducted for two distinct benchmark years: 2010 and 2015. These years were selected because all four emission inventories provide well-documented and internally consistent emissions data for both, enabling a robust inter-comparison.

2.2. Biogenic Emissions

Biomass-burning emissions were taken from the Fire INventory from NCAR (FINN v1.5) [32], which provides global daily estimates of trace-gas and particle emissions from open biomass burning—including wildfires, agricultural fires, and prescribed burns—at 1 km spatial resolution. Monthly biogenic emissions were represented with the Model of Emissions of Gases and Aerosols from Nature (MEGAN v2.1) [33], which estimates net fluxes of isoprene, monoterpenes, and other biogenic trace gases and aerosols from terrestrial ecosystems.
To isolate the effects of anthropogenic emission inventories on simulated pollutant concentrations, biogenic and biomass-burning emissions were derived from a single dataset in each category. This design choice ensures that differences in model outputs arise solely from variations in anthropogenic emissions, allowing for a controlled and systematic comparison among the four inventories.

2.3. Anthropogenic Emissions

We used four widely recognized anthropogenic emissions inventories: the REAS, the CAMS-GLOB-ANT, ECLIPSE, and the HTAP. These inventories provide estimates of emissions for PM2.5, SO2, CO, NOx, and others, each based on distinct methodologies, assumptions, and input datasets.

2.3.1. REAS

The REAS is a comprehensive database covering East, Southeast, and South Asia. The latest release, REAS v3.2.1, spans 1950–2015 and reports SO2, NOx, CO, non-methane volatile organic compounds (NMVOCs), particulate matter (PM10, PM2.5), black carbon (BC), organic carbon (OC), ammonia (NH3), and carbon dioxide (CO2) [9]. Data are provided on a 0.25° × 0.25° grid with monthly resolution, enabling detailed spatio-temporal analysis. Key updates include corrected NMVOC estimates and refined power-plant geolocation, improving spatial and temporal allocation. The inventory covers power generation, industry, transport and the residential sector, plus industrial processes, agriculture (fertilizer application and livestock) and fugitive/solvent sources; international and domestic aviation, maritime navigation, and soil-NOx are excluded. National statistics supply evaporative NMVOC data for Japan and Korea, while NH3 methods follow earlier versions for temporal consistency. REAS v3.2.1 enables long-term, high-resolution analyses of air quality and climate across Asia.

2.3.2. CAMS-GLOB-ANT

CAMS-GLOB-ANT v6.2 is a global inventory developed under the Copernicus Atmosphere Monitoring Service for monthly air quality and climate applications. Covering 2000–2025 at 0.1° × 0.1° resolution, it reports 36 pollutant parameters across 21 sectors. Annual totals from EDGAR v6 are temporally downscaled with CAMS-GLOB-TEMPO profiles, while recent-year trends from the Community Emissions Data System (CEDS) are blended in [7]. Activity data come from globally recognized sources (IEA, FAO, sector-specific databases). Sector-specific temporal profiles and spatial proxies capture regional and seasonal variations more accurately. NMVOC estimation has been substantially refined, making CAMS-GLOB-ANT v6.2 a robust, up-to-date dataset for air quality research and policy evaluation.

2.3.3. ECLIPSE V6b

ECLIPSE v6b is a global inventory produced with the GAINS model under the EU Seventh Framework Programme. It gives emissions for ~200 countries and hundreds of sectors from 1990 to 2030 (five-year steps; monthly grids derived from ECLIPSE v5 profiles) at 0.5° × 0.5° resolution [17]. Pollutants include NOx, SO2, PM2.5, NH3 and VOCs. Recent enhancements provide finer regional detail as well as updated legislation (e.g., China’s 13th Five-Year Plan), a dedicated waste sector, soil-NOx, revised international-shipping estimates, and new spatial patterns for power, flaring, transport and industry. Projections from the IEA World Energy Outlook (2018) and other datasets ensure consistency with future energy and economic pathways. ECLIPSE’s focus on short-lived climate pollutants (SLCPs) facilitates joint air quality–climate impact assessments.

2.3.4. HTAP V3

HTAP v3 is a “mosaic” inventory compiled from multiple regional sources under the UNECE Air Convention and covers 2000–2018. It reports SO2, NOx, CO, NMVOCs, NH3, PM10, PM2.5, BC, and OC on a 0.1° × 0.1° grid with monthly profiles. For Southeast Asia, HTAP v3 incorporates REAS v3.2.1 to preserve regional characteristics [11]. NMVOC emissions follow a refined method [34] that extends the EDGAR framework to provide sector-specific chemical speciation. By integrating regional inventories with detailed speciation, HTAP v3 is suited to transboundary-pollution analysis and to the design of coordinated mitigation strategies.

2.4. Observation Dataset

Observational data for PM2.5, SO2, CO and NOx were obtained from the Air-Pollution Monitoring Data platform of the National Institute for Environmental Studies (NIES), Japan [35]. This platform supplies high-quality, long-term measurements from Japan’s extensive ground network and is essential for validating emission inventories. Hourly, monthly, and annual values are available for major pollutants. The number of monitoring sites in 2010 (2015) is 70 (981) for PM2.5, 942 (898) for SO2, 271 (261) for CO, and 1456 (1488) for NOx (Figures S1–S4). The platform uses standardized protocols to ensure data comparability across sites, and the dataset is openly accessible and regularly updated. In this study, these observations provide the benchmark against which the four selected inventories are evaluated, thereby enhancing the accuracy, reliability and policy relevance of model-based air quality assessments in Japan.

2.5. Data Pre-Processing

All datasets underwent a fully scripted, standardized pre-processing workflow before their use in WRF-Chem simulations and statistical evaluations (Figure 2). The workflow comprised five main steps:
  • Data retrieval: Monthly emissions of PM2.5, SO2, CO, and NOx for 2010 and 2015 were downloaded in ASCII or NetCDF format from the repositories of the four anthropogenic inventories—REAS v3.2.1, CAMS-GLOB-ANT v6.2, ECLIPSE v6b and HTAP v3. Observation data came from the NIES Air-Pollution Monitoring platform. Meteorological boundary conditions were taken from the 0.25° NCEP-FNL analyses; chemical boundaries from the CAM-Chem reanalysis; land-cover from MODIS; and topography from GTOPO30. Natural-source inventories comprised FINN v1.5 (biomass burning) and MEGAN v2.1 (biogenic VOCs).
  • Reprojection: All emissions were imported into GRASS GIS v8.1 and re-projected from their native geographic grids to the Lambert conformal conic 10 km × 10 km projection used by WRF-Chem.
  • Sector aggregation: Inventory-specific sectors—6 in REAS, 11 in CAMS-GLOB-ANT, 10 in ECLIPSE, and 16 in HTAP—were collapsed into four common categories: residential & other, industry, energy, and transportation (see Tables S1–S4). This harmonization enables cross-inventory sectoral comparisons.
  • Unit conversion: Monthly totals (t month−1) were converted to kg m−2 s−1, the unit required by WRF-Chem. Additional species such as NMVOCs and NH3 were included in the simulations but are not evaluated in this paper.
  • Post-processing & evaluation: After each model run, spatial distributions of PM2.5, SO2, CO, and NOx were plotted. Regional mean concentrations were calculated for 2010 and 2015; model skill was assessed with root-mean-square error (RMSE) and mean bias against observations.
This unified workflow guarantees reproducibility and ensures that differences in model performance arise from the underlying emission inventories rather than from data-handling inconsistencies.

2.6. Comparison of Anthropogenic Emission Across Inventories

REAS, CAMS-GLOB-ANT, and HTAP derive emissions with a bottom-up methodology that multiplies activity data by sector-specific emission factors—a standard approach in regional inventories [7,9,11]. In contrast, ECLIPSE is produced with the GAINS model, which couples economic-development pathways, sectoral activity, emission factors, and prospective policy measures [17].
ECLIPSE reports the highest estimated regional annual PM2.5 totals (Table 2). Yearly values exceed 3 × 105 t yr−1 in Kanto (3.40 × 105 t), Chubu (3.42 × 105 t) and Kinki (3.27 × 105 t). CAMS-GLOB-ANT gives the smallest totals (e.g., 1.49 × 105 t yr−1 in Kanto). In every inventory, Kanto, Chubu, and Kinki are the highest-emitting regions, reflecting dense population, industrial clusters, and intense road traffic, whereas Hokkaido and Shikoku are the lowest. ECLIPSE is notable for particularly high residential contributions in Tohoku (1.45 × 105 t), Kanto (1.11 × 105 t), and Chubu (1.38 × 105 t). In Shikoku, ECLIPSE totals 8.27 × 104 t, compared with 3.30 × 104 t (REAS) and 2.65 × 104 t (CAMS-GLOB-ANT). In REAS, 75–92% of PM2.5 originates from industry + transport, whereas HTAP is dominated by transport everywhere except Hokkaido, where residential sources are marginally higher. The energy sector is the smallest PM2.5 contributor in all inventories.
CAMS-GLOB-ANT produces markedly higher SO2 in every region—especially Kanto (3.03 × 106 t yr−1; Table 3). REAS, HTAP, and ECLIPSE yield lower but comparable totals with contrasting sectoral patterns. In REAS, industry is the main SO2 source (e.g., 6.48 × 105 t in Kanto) and transport SO2 is minimal, consistent with Japan’s low-sulfur fuels [36,37]. ECLIPSE assigns substantial SO2 shares to transport, notably 2.24 × 105 t in Kinki and 1.43 × 105 t in Kyushu, underscoring different sectoral allocations.
For CO, CAMS-GLOB-ANT gives the highest CO estimates in six of the eight regions, exceeding 1.47 × 107 t in Kanto and reaching 1.18 × 107 t in Kyushu (Table 4). HTAP is also elevated in Kanto and Chubu. Transport is the dominant CO source in all inventories examined: >9 × 106 t in Kanto for both REAS and HTAP and 5.20 × 106 t for CAMS-GLOB-ANT. ECLIPSE yields the greatest industrial CO in Kinki (8.20 × 106 t) and the highest residential CO in Tohoku (1.04 × 106 t). Energy-sector CO is minor everywhere, though CAMS-GLOB-ANT is comparatively higher in Kanto (1.39 × 106 t) and Kyushu (6.34 × 105 t).
REAS reports the largest NOx totals, led by Kanto (5.92 × 106 t yr−1; Table 5), where transport exceeds half of emissions. HTAP is lower overall but shows a similar sectoral split. CAMS-GLOB-ANT minimizes residential NOx yet maximizes the energy-sector share (e.g., 1.46 × 106 t in Kanto, 5.37 × 105 t in Kyushu). Among the four inventories, ECLIPSE gives the lowest NOx totals across all regions (e.g., 3.27 × 106 t in Kanto) while still identifying transport as the leading source. Kanto, Chubu and Kinki consistently register the highest regional NOx; Hokkaido and Shikoku remain the lowest.

3. Results

3.1. Surface PM2.5 Concentration

Monthly mean PM2.5 over Japan shows a clear west-to-east gradient (Figure 3): the highest 2010 values occurred in Chugoku (18.8 µg m−3), Kinki (16.0 µg m−3), and Shikoku (15.3 µg m−3), reflecting a combination of transboundary inflow and dense local sources. Nationwide concentrations declined between 2010 and 2015, most notably in Kanto (−3.8 µg m−3) and Chugoku (−4.8 µg m−3) (Table 6). Model performance varied significantly depending on the emission inventory used. REAS overestimated PM2.5 in western Japan (+16.0 µg m−3 in Chugoku and +23.7 µg m−3 in Kyushu) while underestimating Tohoku and Shikoku. Although biases narrowed in 2015, RMSEs remained high, particularly in Kyushu (Table 6; Figure 3 and Figures S5–S7). CAMS-GLOB-ANT reproduced the spatial pattern poorly, underpredicting PM2.5 almost everywhere except Hokkaido in 2010, where it showed a small positive bias. ECLIPSE significantly overestimated concentrations in Kanto (+34.2 µg m−3) and Kinki (+35.9 µg m−3) in 2010 but agreed more closely elsewhere; its best RMSE (10.8 µg m−3) occurred in Chugoku. HTAP greatly overpredicted Hokkaido (+48.7 µg m−3, more than 11 times the observed value) and showed a sizeable positive bias in Kanto, yet improved markedly from 2010 to 2015, with Kanto RMSE dropping from 41.3 to 14.3 µg m−3.
In summary, REAS and HTAP tend to overestimate concentrations in western and northern Japan; CAMS-GLOB-ANT generally underestimates nationwide; and ECLIPSE captures the spatial gradient well but significantly overestimates levels in Japan’s two largest metropolitan regions.

3.2. Surface SO2 Concentration

Observed SO2 concentrations were highest in Shikoku (10.7 µg m−3) and other industrialized regions—Kinki, Kyushu, and Chugoku—during 2010 and declined across all regions by 2015 (Table 7). REAS matched observations most closely: in 2010 the mean bias was within ±3 µg m−3 in six of the eight regions, and RMSE values were generally <10 µg m−3, although Shikoku showed a larger negative bias (Table 7; Figure 4 and Figures S8–S10). CAMS-GLOB-ANT exhibited a systematic positive bias, exceeding +40 µg m−3 in Kanto and yielding an RMSE of 134 µg m−3 in 2010; substantial overprediction persisted in 2015. ECLIPSE consistently underestimated concentrations, but only by a few micrograms per cubic meter, and kept RMSEs below 8 µg m−3. HTAP exhibited the most consistent agreement with observations: in 2015, all regional biases lay within ±3 µg m−3 except Hokkaido (+3.5 µg m−3), and RMSEs were comparable to those of REAS. In summary, CAMS-GLOB-ANT markedly overpredicts SO2, ECLIPSE and REAS are generally faithful to observations, and HTAP showed the strongest agreement with observed values.

3.3. Surface CO Concentration

The highest CO concentrations in 2010 were recorded in Kyushu (633 µg m−3) and Kanto (538 µg m−3) (Table 8). REAS underestimated CO in every region, with negative biases ranging from −55% to −90% and RMSE values > 400 µg m−3 in most areas. CAMS-GLOB-ANT showed strong regional dependence: it overpredicted CO concentrations in Kanto (+205 µg m−3, +38%) and Kyushu (+729 µg m−3, +115%) but underpredicted in Hokkaido and Chubu; the Kyushu RMSE exceeded 4300 µg m−3 (Table 8; Figure 5 and Figures S11–S13). ECLIPSE also consistently underestimated CO, with a maximum bias of −563 µg m−3 in Kyushu, though its RMSE values were generally lower than those of REAS. HTAP agreed best with observations, delivering the lowest RMSEs and modest biases; in 2015, six of the eight regions were within ±25% of observed means; Tohoku was the primary outlier, where the bias reached −36%. Consequently, HTAP most accurately reflects observed CO variability, whereas REAS and ECLIPSE consistently underpredict, and CAMS-GLOB-ANT alternates between severe under- and overestimation depending on region.

3.4. Surface NOx Concentration

Observed NOx peaked in Kanto (38.6 µg m−3) and Kinki (32.7 µg m−3) in 2010 (Table 9). REAS greatly overestimated every region, with positive biases of +78.5 µg m−3 in Kanto and +67.2 µg m−3 in Kinki, accompanied by RMSE values of comparable size. CAMS-GLOB-ANT showed mixed skill: it was reasonably accurate in Hokkaido and had a small negative bias in Chubu but overpredicted urban regions such as Kanto and Shikoku, generating very high RMSEs in Kanto and Kinki (Table 9; Figure 6 and Figures S14–S16). ECLIPSE consistently underestimated NOx but produced the smallest RMSEs (20–33 µg m−3) across regions, indicating that it replicates spatial distribution well, despite underestimating overall concentrations. HTAP produced the largest positive biases—> +30 µg m−3 in most regions in 2010—and correspondingly high RMSEs.
In summary, REAS and HTAP strongly overpredict NOx; ECLIPSE underpredicts but reproduces the spatial structure best; and CAMS-GLOB-ANT shows moderate agreement overall, with accuracy differing significantly between rural and urban areas.

4. Discussion

This study benchmarked WRF-Chem simulations driven by four anthropogenic inventories—REAS v3.2.1, CAMS-GLOB-ANT v6.2, ECLIPSE v6b, and HTAP v3—against nationwide observations of PM2.5, SO2, CO, and NOx for 2010 and 2015. Mean bias (MBIAS) and root-mean-square error (RMSE) statistics (Table 6, Table 7, Table 8 and Table 9) reveal distinct, inventory-specific patterns, which are discussed below.

4.1. Systematic over- and Under-Estimation Patterns

REAS generally overestimated PM2.5 and NOx in western Japan (Chugoku, Kyushu); its largest PM2.5 bias (approximately +180%) occurred in Kyushu. Similar REAS-driven overestimations of nitrogen species were reported by Itahashi et al. [38] for the early 2010s, who linked the bias to heavy-duty diesel traffic along the Seto Inland Sea corridor. NOx biases exceeded 200% in the Kanto and Kinki conurbations. Conversely, REAS slightly underpredicted PM2.5 in Shikoku and showed only small positive NOx biases in Tohoku and Shikoku, indicating that industrial and traffic activity may be overestimated along heavily industrialized corridors, whereas small-scale combustion sources remain undervalued elsewhere. CAMS-GLOB-ANT persistently underpredicted PM2.5 (−60% to −85%) and CO and overpredicted SO2, most strikingly in Kanto, where the 2010 bias reached +870%. This SO2 overestimation is consistent with observations from Itahashi et al. [39], who showed discrepancies between satellite SO2 column densities and emission inventories in some parts of Japan, suggesting that assumptions in regional emission factors (including desulfurization) may lead to overestimation under certain conditions. ECLIPSE showed the greatest spatial variability. It overpredicted PM2.5 in Kanto and Kinki by >200% and consistently underpredicted NOx (10–70%). Despite negative NOx biases, ECLIPSE produced the smallest NOx RMSEs, indicating that relative regional gradients are correct while absolute magnitudes are incorrectly scaled—a pattern also documented by Crippa et al. [8] for European cities—possibly because GAINS scenarios weigh short-lived climate pollutants more heavily than conventional criteria gases. HTAP, which integrates regional and global grids, reproduced SO2 and CO concentrations in central Japan accurately (±25% in 2015), but it strongly overpredicted PM2.5 in Hokkaido (~ +1100%) and NOx in several regions (e.g., +120% in Kinki), echoing the mosaic-related misallocations discussed by Li et al. [40] for East Asia, where continental outflow can be double-counted or displaced poleward.

4.2. Regional and Pollutant Contrasts

Over the urban–industrial belts of Kanto, Kinki and Chubu, simulations tended to overestimate NOx (REAS, HTAP) or overestimate PM2.5 (ECLIPSE). Disentangling on-road, non-road, and stationary combustion sources remains difficult because each source type follows a distinct temporal profile and after-treatment penetration rate. Sadanaga et al. [41] similarly showed that region-specific after-treatment ratios are decisive for reproducing NOx diurnal cycles in Osaka. Rural and mountainous areas such as Shikoku exhibited pronounced under-estimation of PM2.5 and CO (REAS, CAMS), indicating that residential heating, small-scale biomass burning, and stable nocturnal boundary layers remain poorly quantified and underrepresented in current inventories. Yamaji et al. [42] reached comparable conclusions for wintertime CO underestimation at remote Japanese sites, attributing the bias to unaccounted agricultural residue burning. Hokkaido presented a contrasting picture: CAMS underpredicted PM2.5, whereas REAS overpredicted it. Western transboundary-influenced regions (Chugoku, Kyushu) showed the largest CAMS-driven CO errors (RMSE > 4300 µg m−3) and REAS-driven PM2.5 errors, underscoring the need for improved estimates of shipping emissions and continental inflow. A recent inverse study by Miyazaki et al. [43] demonstrated that assimilating satellite NOX columns can halve such biases, suggesting a feasible pathway for inventory refinement.

4.3. Implications for Air Quality Modeling and Policy

The inter-inventory spread, which in many cases exceeds a factor of two (i.e., the highest estimate is more than double the lowest), confirms that emissions uncertainty dominates regional chemical-transport modeling errors even when a well-validated dynamical core such as WRF-Chem is used. This conclusion is consistent with the multi-model comparison of Zhang et al. [44], who found that emission variability contributes up to 70% of the PM2.5 simulation spread over East Asia. Because no single inventory performs consistently well across all regions and pollutant types, model-based policy assessments should match inventories to their strengths. For example, HTAP offers the most reliable SO2 and CO baselines in central Japan, whereas ECLIPSE provides the most realistic spatial gradients for NOx, albeit with a negative overall bias. Such tailored usage mirrors the “hybrid-inventory” approach advocated by Kurokawa and Ohara [9] for national emissions reporting in Asia. Relying on a single inventory may therefore misguide policy decisions—overpredicted NOx could falsely magnify expected benefits of emission controls, while underpredicted CO may obscure the incomplete combustion problem in rural prefectures. Regular assimilation of ground-based and satellite observations, together with inverse modeling, is essential for adjusting bottom-up inventories and reducing regional bias.
The fact that identical meteorological fields, chemical mechanisms, and natural-source inputs were used across all simulations yet produced markedly different biases depending on the emission inventory demonstrates that observational mismatches are dominated by emission uncertainty rather than by model physics or measurement errors. For example, in Kanto, NOx concentrations were overestimated by more than +200% with REAS but underestimated by up to −70% with ECLIPSE, despite both being evaluated against the same dense and standardized NIES monitoring network. Such opposite-signed deviations cannot plausibly be explained by observational uncertainty, which is comparatively minor; rather they are explained by divergent assumptions in traffic and industrial emission factors across inventories. Similarly, the severe overestimation of SO2 in Kanto by CAMS-GLOB-ANT (+870% in 2010) and the close agreement of HTAP in central Japan further highlight that the choice of inventory, rather than the observational dataset, is decisive in shaping model–observation discrepancies. This finding corroborates previous multi-model assessments showing that emission variability accounts for the majority of spread in simulated concentrations over East Asia [15], and it reinforces the importance of routine assimilation and inverse modeling to constrain emissions and reduce uncertainty.
These results further demonstrate that policy assessments should adopt pollutant- and region-specific inventories, as no single dataset provides the best performance across all conditions. In addition, ensemble or hybrid-inventory approaches are strongly recommended, as reliance on a single inventory risks misleading policy evaluation. Routine assimilation of ground and satellite observations, such as TROPOMI NO2 columns, represents a promising future pathway for reducing residual biases in emission estimates and improving inventory accuracy.

4.4. Limitations of This Study

This study has at least three key limitations. First, the evaluation relied on monthly mean concentrations, a choice that smooths both diurnal variability and short-lived high-pollution episodes such as Asian dust intrusions or photochemical smog events. Second, the analysis did not treat secondary products—sulfate and secondary organic aerosol formed from SO2 and VOC–NOx chemistry, or ozone produced from NOx—as separate quantities. Omitting these pathways makes it difficult to determine whether errors originate in primary emissions, chemical production, or both. Third, unavoidable uncertainties in meteorological forcing, chemical mechanisms, and the representativeness of monitoring sites all influence the reported bias and RMSE statistics. Future studies should therefore extend the evaluation to high temporal resolution datasets, conduct source-oriented sensitivity experiments that target high-bias sectors such as on-road traffic and residential heating, and integrate WRF-Chem with inverse approaches that assimilate in situ, satellite column, and plume observations to iteratively correct the underlying emission inventories. Another limitation is that sector-specific sensitivity experiments were not conducted. While we harmonized inventory sectors into four categories (residential, industry, energy, transport) for comparability, we did not perform perturbation tests on specific high-bias sectors such as road traffic, residential heating, small-scale biomass burning, or shipping emissions. Such targeted analyses would help quantify the contribution of individual sectors to overall emission uncertainty and should be prioritized in future work. Furthermore, our evaluation did not address secondary pollutants such as sulfate, secondary organic aerosol, or ozone separately. We intentionally restricted our analysis to directly emitted species (PM2.5, SO2, CO, NOx) to avoid conflating errors in emissions with uncertainties in chemical production. Future studies should be expanded to secondary products to provide a more complete assessment of air quality modeling performance.

5. Conclusions

This study benchmarked WRF-Chem simulations driven by four anthropogenic inventories—REAS v3.2.1, CAMS-GLOB-ANT v6.2, ECLIPSE v6b, and HTAP v3—against nationwide observations of PM2.5, SO2, CO, and NOx for 2010 and 2015. ECLIPSE best captured NOx spatial gradients but overestimated metropolitan PM2.5; HTAP yielded the most balanced SO2 and CO fields yet overstated PM2.5 in Hokkaido and urban NOx. REAS reproduced industrial SO2 but overstated western PM2.5 and NOx, whereas CAMS-GLOB-ANT consistently underpredicted PM2.5/CO and overpredicted SO2. Inter-inventory spreads exceeding a factor of two confirm that emissions uncertainty, not model physics, dominates regional air quality error. Inventory choice should therefore be pollutant- and region-specific, and ensemble approaches are recommended for policy analysis. Routine assimilation of ground and satellite data, coupled with inverse modeling, offers a practical route to narrow bias. Future research should evaluate diurnal extremes, secondary aerosol and ozone, and perform sector-targeted sensitivity tests to meet emerging precision demands.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/data10090151/s1, Figure S1: Distribution of monthly average PM2.5 concentration using observed station anthropogenic emissions for 2015. at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S2: Distribution of monthly average SO2 concentration using Observed station anthropogenic emissions for 2015. at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S3: Distribution of monthly average CO concentration using Observed station anthropogenic emissions for 2015. at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S4: Distribution of monthly average NOx concentration using Observed station anthropogenic emissions for 2015. at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S5: Distribution of monthly average PM2.5 concentration using CAMS-GLOB-ANT anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S6: Distribution of monthly average PM2.5 concentration using ECLIPSE anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S7: Distribution of monthly average PM2.5 concentration using HTAP anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S8: Distribution of monthly average SO2 concentration using CAMS-GLOB-ANT anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S9: Distribution of monthly average SO2 concentration using ECLIPSE anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S10: Distribution of monthly average SO2 concentration using HTAP anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S11: Distribution of monthly average CO concentration using CAMS-GLOB-ANT anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S12: Distribution of monthly average CO concentration using ECLIPSE anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S13: S Distribution of monthly average CO concentration using HTAP anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S14: Distribution of monthly average NOx concentration using CAMS-GLOB-ANT anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S15: Distribution of monthly average NOx concentration using ECLIPSE anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Figure S16: Distribution of monthly average NOx concentration using HTAP anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December; Table S1: Mapping of REAS original inventory sectors to major sector categories; Table S2: Mapping of CAMS-GLOB-ANT original inventory sectors to major sector categories; Table S3: Mapping of ECLIPSE original inventory sectors to major sector categories; Table S4: Mapping of HTAP original inventory sectors to major sector categories.

Author Contributions

Conceptualization, K.T.; methodology, K.T.; validation, K.T.; formal analysis, K.T.; investigation, K.T.; resources, K.T.; data curation, K.T.; writing—original draft preparation, K.T.; writing—review and editing, K.T. and N.T.H.D.; visualization, K.T.; supervision, N.T.H.D.; project administration, K.T.; funding acquisition, K.T. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by JST PRESTO, grant number JPMJPR16O3 and JSPS KA-KENHI, grant number 16KK0169 and 19K15944.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are available on request from the author.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
Pollutants & Compounds
PM2.5Fine Particulate Matter (particles with diameter ≤ 2.5 μm)
SO2Sulfur Dioxide
NOxNitrogen Oxides (includes NO and NO2)
COCarbon Monoxide
HONONitrous Acid
Emission Inventories
REASRegional Emission inventory in Asia
CAMS-GLOB-ANTCopernicus Atmosphere Monitoring Service Global Anthropogenic Emissions
ECLIPSEEvaluating the Climate and Air Quality Impacts of Short-Lived Pollutants
HTAPHemispheric Transport of Air Pollution
Models & Datasets
WRF-ChemWeather Research and Forecasting Model coupled with Chemistry
MODISModerate Resolution Imaging Spectroradiometer
GTOPO30Global 30 Arc-Second Elevation Dataset
NCEP FNLNational Centers for Environmental Prediction Final Analysis
CAM-ChemCommunity Atmosphere Model with Chemistry
MOZART-4Model for Ozone and Related Chemical Tracers, version 4
MOSAICModel for Simulating Aerosol Interactions and Chemistry
GOCARTGoddard Chemistry Aerosol Radiation and Transport
FINNFire INventory from NCAR
MEGANModel of Emissions of Gases and Aerosols from Nature
Other Terms
RMSERoot Mean Square Error
NCARNational Center for Atmospheric Research

References

  1. World Health Organization. WHO Global Air Quality Guidelines: Particulate Matter (PM2.5 and PM10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide; World Health Organization: Geneva, Switzerland, 2021; Available online: https://www.who.int/publications/i/item/9789240034228 (accessed on 5 February 2025).
  2. Lelieveld, J.; Evans, J.S.; Fnais, M.; Giannadaki, D.; Pozzer, A. The contribution of outdoor air-pollution sources to premature mortality on a global scale. Nature 2015, 525, 367–371. [Google Scholar] [CrossRef]
  3. Landrigan, P.J.; Fuller, R.; Acosta, N.J.R.; Adeyi, O.; Arnold, R.; Baldé, A.B.; Bertollini, R.; Bose-O’Reilly, S.; Boufford, J.I.; Breysse, P.N.; et al. The Lancet Commission on pollution and health. Lancet 2018, 391, 462–512. [Google Scholar] [CrossRef]
  4. Pope, C.A., III; Dockery, D.W. Health effects of fine particulate air pollution: Lines that connect. J. Air Waste Manag. Assoc. 2006, 56, 709–742. [Google Scholar] [CrossRef] [PubMed]
  5. Ministry of the Environment, Japan. Annual Report on the Environment, the Sound Material-Cycle Society and Biodiversity in Japan 2023; Ministry of the Environment: Tokyo, Japan, 2023; Available online: https://www.env.go.jp/content/000175165.pdf (accessed on 12 January 2025).
  6. Zhang, Q.; Jiang, X.; Tong, D.; Davis, S.J.; Zhao, H.; Geng, G.; Feng, T.; Zheng, B.; Lu, Z.; Streets, D.G.; et al. Transboundary health impacts of transported global air pollution and international trade. Nature 2017, 543, 705–709. [Google Scholar] [CrossRef]
  7. Soulie, A.; Granier, C.; Darras, S.; Zilbermann, N.; Doumbia, T.; Guevara, M.; Jalkanen, J.-P.; Keita, S.; Liousse, C.; Crippa, M.; et al. Global anthropogenic emissions (CAMS-GLOB-ANT) for the Copernicus Atmosphere Monitoring Service simulations of air quality forecasts and reanalyses. Earth Syst. Sci. Data 2024, 16, 2261–2279. [Google Scholar] [CrossRef]
  8. Crippa, M.; Solazzo, E.; Huang, G.; Guizzardi, D.; Koffi, E.; Muntean, M.; Schieberle, C.; Friedrich, R.; Janssens-Maenhout, G. High-resolution temporal profiles in the Emissions Database for Global Atmospheric Research. Sci. Data 2020, 7, 121. [Google Scholar] [CrossRef] [PubMed]
  9. Kurokawa, J.; Ohara, T. Long-term historical trends in air-pollutant emissions in Asia: REAS version 3. Atmos. Chem. Phys. 2020, 20, 12761–12793. [Google Scholar] [CrossRef]
  10. Li, M.; Zhang, Q.; Kurokawa, J.I.; Woo, J.H.; He, K.; Lu, Z.; Ohara, T.; Song, Y.; Streets, D.G.; Carmichael, G.R.; et al. MIX: A mosaic Asian anthropogenic emission inventory. Atmos. Chem. Phys. 2017, 17, 935–963. [Google Scholar] [CrossRef]
  11. Janssens-Maenhout, G.; Crippa, M.; Guizzardi, D.; Dentener, F.; Muntean, M.; Pouliot, G.; Keating, T.; Zhang, Q.; Kurokawa, J.; Wankmüller, R.; et al. HTAP_v2.2: A mosaic of regional and global emission grid maps. Atmos. Chem. Phys. 2015, 15, 11411–11432. [Google Scholar] [CrossRef]
  12. Thunis, P.; Crippa, M.; Cuvelier, C.; Guizzardi, D.; de Meij, A.; Oreggioni, G.; Pisoni, E. Sensitivity of air quality modelling to different emission inventories: A case study over Europe. Atmos. Environ. X 2021, 10, 100111. [Google Scholar] [CrossRef]
  13. Guttikunda, S.K.; Goel, R.; Pant, P. Nature of air pollution, emission sources, and management in Indian cities. Atmos. Environ. 2014, 95, 501–510. [Google Scholar] [CrossRef]
  14. Stohl, A.; Aamaas, B.; Amann, M.; Baker, L.H.; Bellouin, N.; Berntsen, T.K.; Boucher, O.; Cherian, R.; Collins, W.; Daskalakis, N.; et al. Evaluating the climate and air quality impacts of short-lived pollutants. Atmos. Chem. Phys. 2015, 15, 10529–10566. [Google Scholar] [CrossRef]
  15. Nishitateno, S.; Burke, P.J.; Arimura, T.H. Road traffic flow and air pollution concentrations: Evidence from Japan. Int. J. Econ. Policy Stud. 2024, 18, 357–385. [Google Scholar] [CrossRef]
  16. Qu, Z.; Henze, D.K.; Worden, H.M.; Jiang, Z.; Gaubert, B.; Theys, N.; Wang, W. Sector-Based Top-Down Estimates of NOx, SO2, and CO Emissions in East Asia (2005–2012). Geophys. Res. Lett. 2022, 49, e2021GL096009. [Google Scholar] [CrossRef]
  17. Klimont, Z.; Kupiainen, K.; Heyes, C.; Purohit, P.; Cofala, J.; Rafaj, P.; Borken-Kleefeld, J.; Schöpp, W. Global anthropogenic emissions of particulate matter including black carbon. Atmos. Chem. Phys. 2017, 17, 8681–8723. [Google Scholar] [CrossRef]
  18. Grell, G.A.; Freitas, S.R.; Stuefer, M.; Fast, J.D. Inclusion of biomass burning in WRF-Chem: Impact of wildfires on weather forecasts. Atmos. Chem. Phys. 2011, 11, 5289–5303. [Google Scholar] [CrossRef]
  19. Grell, G.A.; Peckham, S.E.; Schmitz, R.; McKeen, S.A.; Frost, G.; Skamarock, W.C.; Eder, B. Fully coupled “online” chemistry within the WRF model. Atmos. Environ. 2005, 39, 6957–6975. [Google Scholar] [CrossRef]
  20. National Center for Atmospheric Research (NCAR). NCEP FNL Operational Model Global Tropospheric Analyses, Continuing from July 1999. Research Data Archive. Available online: https://doi.org/10.5065/D6M043C6 (accessed on 12 March 2025).
  21. Lamarque, J.-F.; Emmons, L.K.; Hess, P.G.; Kinnison, D.E.; Tilmes, S.; Vitt, F.; Heald, C.L.; Holland, E.A.; Lauritzen, P.H.; Neu, J.L.; et al. CAM-chem: Description and Evaluation of Interactive Atmospheric Chemistry in the Community Earth System Model. Geosci. Model Dev. 2012, 5, 369–411. [Google Scholar] [CrossRef]
  22. Emmons, L.K.; Walters, S.; Hess, P.G.; Lamarque, J.-F.; Pfister, G.G.; Fillmore, D.; Granier, C.; Guenther, A.; Kinnison, D.; Laepple, T.; et al. Description and evaluation of the Model for Ozone and Related chemical Tracers, version 4 (MOZART-4). Geosci. Model Dev. 2010, 3, 43–67. [Google Scholar] [CrossRef]
  23. Zaveri, R.A.; Easter, R.C.; Fast, J.D.; Peters, L.K. Model for Simulating Aerosol Interactions and Chemistry (MOSAIC). J. Geophys. Res. 2008, 113, D13204. [Google Scholar] [CrossRef]
  24. Ginoux, P.; Chin, M.; Tegen, I.; Prospero, J.M.; Holben, B.; Dubovik, O.; Lin, S.J. Sources and distributions of dust aerosols simulated with the GOCART model. J. Geophys. Res. 2001, 106, 20255–20273. [Google Scholar] [CrossRef]
  25. Morrison, H.; Curry, J.A.; Khvorostyanov, V.I. A new double-moment microphysics parameterization for application in cloud and climate models. Part I: Description. J. Atmos. Sci. 2005, 62, 1665–1677. [Google Scholar] [CrossRef]
  26. Grell, G.A.; Freitas, S.R. A scale- and aerosol-aware stochastic convective parameterization for weather and air-quality modeling. Atmos. Chem. Phys. 2014, 14, 5233–5250. [Google Scholar] [CrossRef]
  27. Jiménez, P.A.; Dudhia, J.; González-Rouco, J.F.; Navarro, J.; Montávez, J.P.; García-Bustamante, E. A revised scheme for the WRF surface-layer formulation. Mon. Weather Rev. 2012, 140, 898–918. [Google Scholar] [CrossRef]
  28. Tewari, M.; Chen, F.; Wang, W.; Dudhia, J.; LeMone, M.A.; Mitchell, K.; Ek, M.; Gayno, G.; Wegiel, J.; Cuenca, R. Implementation and verification of the unified Noah land-surface model in the WRF model. In Proceedings of the 20th Conference on Weather Analysis and Forecasting/16th Conference on Numerical Weather Prediction, Seattle, WA, USA, 11–15 January 2004; American Meteorological Society: Seattle, WA, USA, 2004. Paper 14.2A. [Google Scholar]
  29. Hong, S.Y.; Noh, Y.; Dudhia, J. A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Weather Rev. 2006, 134, 2318–2341. [Google Scholar] [CrossRef]
  30. Mlawer, E.J.; Taubman, S.J.; Brown, P.D.; Iacono, M.J.; Clough, S.A. Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res. 1997, 102, 16663–16682. [Google Scholar] [CrossRef]
  31. Wesely, M.L. Parameterization of surface resistances to gaseous dry deposition in regional-scale numerical models. Atmos. Environ. 1989, 23, 1293–1304. [Google Scholar] [CrossRef]
  32. Wiedinmyer, C.; Akagi, S.K.; Yokelson, R.J.; Emmons, L.K.; Al-Saadi, J.A.; Orlando, J.J.; Soja, A. The Fire INventory from NCAR (FINN): A high-resolution global model to estimate the emissions from open burning. Geosci. Model Dev. 2011, 4, 625–641. [Google Scholar] [CrossRef]
  33. Guenther, A.B.; Jiang, X.; Heald, C.L.; Sakulyanontvittaya, T.; Duhl, T.; Emmons, L.K.; Wang, X. The Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN 2.1). Geosci. Model Dev. 2012, 5, 1471–1492. [Google Scholar] [CrossRef]
  34. Huang, G.; Brook, R.; Crippa, M.; Janssens-Maenhout, G.; Schieberle, C.; Dore, C.; Guizzardi, D.; Muntean, M.; Schaaf, E.; Friedrich, R. Speciation of anthropogenic emissions of non-methane VOCs: A global gridded dataset for 1970–2012. Atmos. Chem. Phys. 2017, 17, 7683–7701. [Google Scholar] [CrossRef]
  35. National Institute for Environmental Studies (NIES). Atmospheric Environmental Regional Observation System (AEROS)/Air-Pollution Monitoring Data Download Service. Available online: https://tenbou.nies.go.jp/download/ (accessed on 30 January 2025).
  36. Central Environment Council (CEC). Future Policy for Motor Vehicle Emission Reduction (Seventh Report); Ministry of the Environment: Tokyo, Japan, 2003; Available online: https://www.env.go.jp/en/air/aq/mv/vehicle-7th.pdf (accessed on 5 May 2025).
  37. TransportPolicy.net. Japan: Fuels—Diesel and Gasoline. Available online: https://www.transportpolicy.net/standard/japan-fuels-diesel-and-gasoline/ (accessed on 5 May 2025).
  38. Itahashi, S.; Yamaji, K.; Chatani, S.; Kitayama, K.; Morino, Y.; Nagashima, T.; Saito, M.; Takigawa, M.; Morikawa, T.; Kanda, I.; et al. Model Performance Differences in Fine-Mode Nitrate Aerosol during Wintertime over Japan in the J-STREAM Model Inter-Comparison Study. Atmosphere 2020, 11, 511. [Google Scholar] [CrossRef]
  39. Itahashi, S.; Hatakeyama, S.; Shimada, K.; Takami, A. Sources of High Sulfate Aerosol Concentration Observed at Cape Hedo in Spring 2012. Aerosol and Air Qual. Res. 2019, 19, 587–600. [Google Scholar] [CrossRef]
  40. Li, M.; Liu, H.; Geng, G.; Hong, C.; Liu, F.; Song, Y.; Tong, D.; Zheng, B.; Cui, H.; Man, H.; et al. Anthropogenic Emission Inventories in China: A Review. Natl. Sci. Rev. 2017, 4, 834–866. [Google Scholar] [CrossRef]
  41. Sadanaga, Y.; Shibata, S.; Hamana, M.; Takenaka, N.; Bandow, H. Weekday/Weekend Difference of Ozone and Its Precursors in Urban Areas of Japan, Focusing on Nitrogen Oxides and Hydrocarbons. Atmos. Environ. 2008, 42, 4708–4723. [Google Scholar] [CrossRef]
  42. Yamaji, K.; Li, J.; Uno, I.; Kanaya, Y.; Irie, H.; Takigawa, M.; Komazaki, Y.; Pochanart, P.; Liu, Y.; Tanimoto, H.; et al. Impact of open crop residual burning on air quality over Central Eastern China during the Mount Tai Experiment 2006 (MTX2006). Atmos. Chem. Phys. 2010, 10, 7353–7368. [Google Scholar] [CrossRef]
  43. Miyazaki, K.; Eskes, H.; Sudo, K.; Boersma, K.F.; Bowman, K.; Kanaya, Y. Decadal changes in global surface NOx emissions from multi-constituent satellite data assimilation. Atmos. Chem. Phys. 2017, 17, 807–820. [Google Scholar] [CrossRef]
  44. Zhang, Q.; Zheng, Y.; Tong, D.; Shao, M.; Wang, S.; Zhang, Y.; Xu, X.; Wang, J.; He, H.; Liu, W.; et al. Drivers of improved PM2.5 air quality in China. Proc. Natl. Acad. Sci. USA 2019, 116, 24463–24469. [Google Scholar] [CrossRef]
Data 10 00151 g001
Figure 1. Regions of Japan.
Figure 1. Regions of Japan.
Data 10 00151 g001
Data 10 00151 g002
Figure 2. Flowchart used to estimate anthropogenic emission concentration.
Figure 2. Flowchart used to estimate anthropogenic emission concentration.
Data 10 00151 g002
Data 10 00151 g003
Figure 3. Distribution of monthly average PM2.5 concentration (μg m−3) using REAS anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Figure 3. Distribution of monthly average PM2.5 concentration (μg m−3) using REAS anthropogenic emissions for 2015 at (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Data 10 00151 g003
Data 10 00151 g004aData 10 00151 g004b
Figure 4. Distribution of monthly average SO2 concentration (μg m−3) using REAS anthropogenic emissions for 2015 in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Figure 4. Distribution of monthly average SO2 concentration (μg m−3) using REAS anthropogenic emissions for 2015 in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Data 10 00151 g004aData 10 00151 g004b
Data 10 00151 g005aData 10 00151 g005b
Figure 5. Distribution of monthly average CO concentration (μg m−3) using REAS anthropogenic emissions for 2015 in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Figure 5. Distribution of monthly average CO concentration (μg m−3) using REAS anthropogenic emissions for 2015 in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Data 10 00151 g005aData 10 00151 g005b
Data 10 00151 g006
Figure 6. Distribution of monthly average NOx concentration (μg m−3) using REAS anthropogenic emissions for 2015 in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Figure 6. Distribution of monthly average NOx concentration (μg m−3) using REAS anthropogenic emissions for 2015 in (a) January, (b) February, (c) March, (d) April, (e) May, (f) June, (g) July, (h) August, (i) September, (j) October, (k) November, and (l) December.
Data 10 00151 g006
Table 1. Chemical scheme used in this study.
Table 1. Chemical scheme used in this study.
SettingSchemesReferences
Chemistry and aerosols schemeMOZART-MOSAIC[22,23]
Dust aerosolsGOCART[24]
MicrophysicsMorrison double-moment[25]
Cumulus parameterizationNew Grell 3D[26]
Surface layerMM5 Monin-obukhov[27]
Land surface modelUnified Noah land-surface model[28]
Planetary boundary layerYSU[29]
Long- and short-wave radiationRRTMG[30]
DepositionWesely[31]
Table 2. The PM2.5 emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
Table 2. The PM2.5 emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
InventoriesSectorsHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushu
REASResidential and other sectors679220,20838,80823,79620,8208508357614,040
Industry22,57230,084104,46067,40470,50098,86813,524135,924
Energy44404452564045844344672011763288
Transportation29,19649,728106,80083,92871,74831,02014,71250,508
TOTAL63,000104,472255,708179,712167,412145,11632,988203,760
CAMS-GLOB-ANTResidential and other sectors11,18415,82851,16830,61228,04410,020535220,952
Industry379216,22444,92818,43232,01612,288460816,680
Energy12122640975630244224181219325568
Transportation19,29636,42042,78054,73234,26026,91614,65236,504
TOTAL35,48471,112148,632106,80098,54451,03626,54479,704
ECLIPSEResidential and other sectors38,808144,732111,156138,31265,04035,53220,568104,796
Industry11,83226,84497,76488,140148,71633,48024,50419,080
Energy4584932421,49213,02026,41214,280660025,788
Transportation26,71254,888109,068102,55286,73636,42030,94866,204
TOTAL81,936235,788339,480342,024326,904119,71282,620215,868
HTAPResidential and other sectors41,35238,12448,55239,52824,72012,960690027,912
Industry860414,40036,13234,11625,56019,872939617,832
Energy1692259216321512984291620523756
Transportation40,05664,284130,104125,85698,06452,53633,40898,112
TOTAL91,704119,400216,420201,012149,32888,28451,756147,612
Table 3. The SO2 emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
Table 3. The SO2 emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
InventoriesSectorsHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushu
REASResidential and other sectors35,96493,612277,644152,244151,89656,66422,068101,292
Industry57,996149,340648,168421,692406,320669,888163,056755,076
Energy167,48481,348209,988104,604220,152229,58457,42090,708
Transportation5016904819,62015,32412,768542426409240
TOTAL266,460333,3481,155,420693,864791,136961,560245,184956,316
CAMS-GLOB-ANTResidential and other sectors91,248171,840589,464330,300344,796116,10064,344227,244
Industry168,072601,3081,664,880766,5601,228,512546,024205,368611,496
Energy185,412271,968747,348248,220328,596203,904230,280636,804
Transportation12,92414,72427,64830,73231,99219,47629,80840,884
TOTAL457,6561,059,8403,029,3401,375,8121,933,896885,504529,8001,516,428
ECLIPSEResidential and other sectors21,72060,43287,84075,54066,66050,96416,27248,276
Industry37,644150,240470,676790,944436,992235,296228,096130,128
Energy51,372101,448230,268137,232194,328123,39662,208283,692
Transportation17,77241,05287,57683,652223,78864,740152,112143,160
TOTAL128,508353,172876,3601,087,368921,768474,396458,688605,256
HTAPResidential and other sectors73,236114,420258,300184,440140,10055,84836,132119,040
Industry105,228137,760377,220331,380291,336286,536128,472212,508
Energy181,476103,57258,28499,00072,132181,66869,084225,060
Transportation29,49629,78490,57674,05279,32069,85274,616156,552
TOTAL389,436385,536784,380688,872582,888593,904308,304713,160
Table 4. The CO emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
Table 4. The CO emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
InventoriesSectorsHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushu
REASResidential and other sectors253,884589,4881,331,508822,576685,056283,956154,464501,504
Industry785,244657,6603,371,1242,277,4682,381,1722,908,908635,6281,619,844
Energy134,72464,728223,404127,584175,776163,53633,90067,692
Transportation1,336,6443,080,4369,065,6766,008,9285,129,4842,040,8041,083,7203,786,204
TOTAL2,510,4964,392,3121,399,17129,236,5568,371,4885,397,2041,907,7125,975,244
CAMS-GLOB-ANTResidential and other sectors306,888531,4201,477,728968,328819,936328,452168,864657,324
Industry84,840364,8966,677,304430,5482,742,300267,252102,9245,615,712
Energy36,75686,9281,389,108199,908423,45691,63260,804633,528
Transportation2,324,9524,873,4525,199,9486,786,8523,897,2163,238,5481,769,6404,932,252
TOTAL2,753,4365,856,69614,744,0888,385,6367,882,9083,925,8842,102,23211,838,816
ECLIPSEResidential and other sectors240,9481,035,444603,372724,560376,044189,27683,952534,732
Industry430,1881,255,2244,125,7923,371,8208,197,884839,0641,185,000695,628
Energy63,744150,312303,840255,972270,216127,11666,516244,392
Transportation892,2122,007,7445,700,4444,109,7482,641,0561,265,820575,1242,163,684
TOTAL1,627,0924,448,72410,733,4488,462,10011,485,2002,421,2761,910,5923,638,436
HTAPResidential and other sectors319,788263,544621,228386,244112,968112,96865,208270,456
Industry625,572861,7202,270,0641,789,3201,617,2401,494,924681,1561,619,148
Energy117,624183,22826,40085,51216,920143,44865,376204,384
Transportation1,746,8763,970,9929,236,6407,591,0565,895,6602,574,6361,488,8885,328,672
TOTAL2,809,8605,279,48412,154,3329,852,1327,642,7884,325,9762,300,6287,422,660
Table 5. The NOx emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
Table 5. The NOx emissions (t yr−1) totals from four anthropogenic inventories, and their sectoral contributions from residential and other sectors, industry, energy, and transportation.
InventoriesSectorsHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushu
REASResidential and other sectors97,800277,704769,140429,384390,744165,06073,260275,856
Industry225,024376,5961,798,428931,752998,5681,051,656153,564756,060
Energy219,540154,992649,236307,356425,400410,74877,916192,204
Transportation620,7721,134,4322,707,4042,009,7721,749,528710,028360,6721,197,372
TOTAL1,163,1361,943,7245,924,2083,678,2643,564,2402,337,492665,4122,421,492
CAMS-GLOB-ANTResidential and other sectors93,624130,164475,044253,596258,52881,13245,720170,100
Industry96,312294,6001,026,348415,728717,588357,108127,524408,744
Energy150,264328,9801,457,244506,604631,656258,132241,296536,880
Transportation346,428637,224760,128975,648631,668493,416280,260654,732
TOTAL686,6281,390,9683,718,7642,151,5762,239,4401,189,788694,8001,770,456
ECLIPSEResidential and other sectors124,464260,652395,784280,416217,812130,51255,908175,272
Industry68,256208,332915,456805,476572,136415,752125,256199,680
Energy95,040164,916442,728346,092373,776153,88866,480417,852
Transportation292,548613,0321,514,3161,201,8361,082,712435,468403,644801,120
TOTAL580,3081,246,9323,268,2842,633,8202,246,4361,135,620651,2881,593,924
HTAPResidential and other sectors236,088293,916737,640472,320395,100143,89285,716303,156
Industry238,920321,732909,048760,860723,960588,348279,660556,056
Energy125,892260,976160,356183,540111,732232,34496,564276,360
Transportation551,388937,8362,195,9041,805,0161,423,428688,188439,3201,299,336
TOTAL1,152,2881,814,4604,002,9483,221,7362,654,2201,652,772901,2602,434,908
Table 6. Statistical metric evaluation (in units of μg m−3) for comparing PM2.5 from WRF-Chem and observations for 2010 and 2015.
Table 6. Statistical metric evaluation (in units of μg m−3) for comparing PM2.5 from WRF-Chem and observations for 2010 and 2015.
InventoriesMetricsYearHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushuTotal
ObservedMean20104.3611.6316.0612.9315.9718.8315.2713.2214.55
Mean20159.0310.9112.2611.7812.514.0314.1815.1812.80
REASMean201013.34.0122.37.922.2734.874.3536.9118.42
RMSE9.928.8310.913.0217.6326.7611.3747.2820.77
MBIAS 18.94−7.626.24−5.036.316.05−10.9223.683.87
Mean201510.092.8112.456.7610.915.292.8214.3910.28
RMSE12.549.329.2811.3811.3318.9712.1426.9715.59
MBIAS1.06−8.10.19−5.01−1.611.26−11.36−0.79−2.52
CAMS-GLOB-ANTMean20109.966.816.775.1611.593.053.610.139.33
RMSE6.769.612.9810.115.2916.7112.1619.8713.72
MBIAS5.61−4.830.71−7.77−4.37−15.77−11.66−3.1−5.22
Mean20154.412.666.283.474.952.842.192.784.05
RMSE5.959.569.59.5410.3712.4812.713.4810.86
MBIAS−4.62−8.25−5.98−8.3−7.55−11.19−11.99−12.4−8.76
ECLIPSEMean201011.75750.2915.6551.912.0415.2215.7229.80
RMSE9.26.9341.2618.3653.2210.828.4914.733.45
MBIAS7.39−4.6334.232.7235.93−6.79−0.052.515.25
Mean20157.368.0225.9616.8133.518.2311.29.9318.50
RMSE4.947.6423.4218.7637.529.28.738.221.38
MBIAS−1.67−2.913.75.0421.01−5.8−2.98−5.255.70
HTAPMean201053.0512.1235.612.2422.7810.6823.4531.9922.25
RMSE49.368.5941.288.2717.9910.8617.1131.3224.12
MBIAS48.70.4919.54−0.686.81−8.158.1918.777.71
Mean2015156.1311.899.19.828.819.297.519.42
RMSE12.47.9214.289.8311.0910.4913.4411.9611.72
MBIAS5.97−4.78−0.37−2.68−2.68−5.22−4.89−7.67−3.38
1 Mean bias.
Table 7. Statistical metric evaluation (in units of μg m−3) for comparing SO2 from WRF-Chem and observations for 2010 and 2015.
Table 7. Statistical metric evaluation (in units of μg m−3) for comparing SO2 from WRF-Chem and observations for 2010 and 2015.
InventoriesMetricsYearHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushuTotal
ObservedMean20106.543.075.115.076.976.110.747.555.93
Mean20153.572.193.594.184.835.927.835.874.57
REASMean20103.091.395.922.398.277.953.337.335.22
RMSE4.963.727.096.7810.8410.199.9613.659.06
MBIAS−3.45−1.690.82−2.681.31.85−7.41−0.22−0.71
Mean20153.161.296.632.256.457.872.716.525.00
RMSE4.223.278.635.957.8210.177.8912.428.55
MBIAS−0.41−0.93.04−1.931.611.95−5.110.650.42
CAMS-GLOB-ANTMean201021.917.3549.7414.4742.5623.6331.3124.0829.99
RMSE50.2137.8134.4623.1998.7260.251.6351.9881.64
MBIAS15.3714.2744.639.435.5917.5320.5716.5324.05
Mean201521.0214.6942.0317.8535.2322.6224.4320.6426.61
RMSE50.0732.99102.5938.0281.1154.4544.450.6267.06
MBIAS17.4512.538.4413.6630.3916.7116.614.7822.03
ECLIPSEMean20101.480.845.034.256.872.643.752.153.81
RMSE6.393.75.827.166.715.569.758.626.78
MBIAS−5.06−2.23−0.08−0.82−0.11−3.46−6.98−5.4−2.12
Mean20151.310.724.24.085.962.413.541.793.32
RMSE3.842.984.956.656.35.747.487.76.08
MBIAS−2.26−1.470.61−0.11.13−3.5−4.29−4.07−1.25
HTAPMean20107.114.754.453.315.177.617.794.194.84
RMSE11.6610.3110.357.337.9811.3313.0611.5610.03
MBIAS0.581.67−0.65−1.76−1.81.51−2.95−3.37−1.09
Mean20157.084.353.963.244.247.245.863.454.32
RMSE13.411.229.57.197.0111.2412.5610.169.71
MBIAS3.512.160.37−0.94−0.591.33−1.97−2.42−0.25
Table 8. Statistical metric evaluation (in units of μg m−3) for comparing CO from WRF-Chem and observations for 2010 and 2015.
Table 8. Statistical metric evaluation (in units of μg m−3) for comparing CO from WRF-Chem and observations for 2010 and 2015.
InventoriesMetricsYearHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushuTotal
ObservedMean2010466.31423.76537.54513.11533.65521.31462.96633.36527.70
Mean2015385.55343415.38399.81457.66412.54477.95494.62422.85
REASMean2010216.85121.13650.69330.33586.17422.594258.36459.61
RMSE401.55357.06398.15480.46445.32407.96389.31493.44429.43
MBIAS−249.47−302.63113.16−182.7952.52−98.81−368.96−375−68.09
Mean2015189.295.37482.22251.58402.37382.9370.02216.95347.96
RMSE330.78279.81340.3359.55326.33410.02420.91420.12353.60
MBIAS−196.35−247.6366.84−148.23−55.29−29.61−407.93−277.67−74.89
CAMS-GLOB-ANTMean2010273.11261.54742.82296.58386.58210.69187.641362.08553.46
RMSE317.95216.091843.62309.12335.25377.08300.924315.721728.13
MBIAS−193.2−162.22205.28−216.54−147.06−310.62−275.32728.7225.76
Mean2015205.61240.98671.65277.42334.2187.8186.921376.83508.85
RMSE290.48150.941726.33214.78262.49296.13325.294337.511656.65
MBIAS−179.94−102.02256.27−122.39−123.46−224.74−291.03882.2186.00
ECLIPSEMean201045.4453.26309.81176.34535.0278.99120.4270.29256.83
RMSE449.43410.9327.25442.69483.31473.05394.21604.12432.30
MBIAS−420.87−370.5−227.73−336.781.37−442.32−342.54−563.07−270.87
Mean201541.9747.59250.16154.54481.3567.8567.968.87218.84
RMSE372.79325.04252.5341.56437.64385.27436.73491.61356.32
MBIAS−343.59−295.41−165.22−245.2723.69−344.69−410.05−425.75−204.01
HTAPMean2010580.39310.26638.8477.71658.77462.8519.98641.54572.95
RMSE478.09186.26396.03314.82537.9298.12329.35442.25401.38
MBIAS114.08−113.51101.26−35.4125.12−58.5157.028.1745.25
Mean2015470.98222.36436.55310.63417.76305.87456.84469.73389.22
RMSE391.7172.58286.16212.11378.82269.41317.2395.74300.87
MBIAS85.43−120.6521.17−89.19−39.9−106.67−21.1−24.89−33.64
Table 9. Statistical metric evaluation (in units of μg m−3) for comparing NOx from WRF-Chem and observations for 2010 and 2015.
Table 9. Statistical metric evaluation (in units of μg m−3) for comparing NOx from WRF-Chem and observations for 2010 and 2015.
InventoriesMetricsYearHokkaidoTohokuKantoChubuKinkiChugokuShikokuKyushuTotal
ObservedMean201026.720.238.6428.532.6527.4823.6423.129.56
Mean201522.5415.6628.7821.6324.721.0416.7117.8722.33
REASMean201036.8822.59117.1854.6599.7962.7725.7342.469.98
RMSE35.4519.33118.3370.8491.1976.1821.4653.4881.24
MBIAS10.182.478.5326.1467.1535.292.0919.3140.42
Mean201532.4717.9498.5443.2375.3558.418.235.5956.73
RMSE32.717.02110.2858.1169.5176.4816.4347.9271.50
MBIAS9.932.2869.7721.650.6537.361.4817.7234.40
CAMS-GLOB-ANTMean201025.2721.4960.7223.3149.3624.9541.227.5537.22
RMSE53.9539.42238.5934.399.8253.2157.647.47127.40
MBIAS−1.431.322.08−5.1916.71−2.5317.564.457.66
Mean201524.219.9452.4928.3244.6624.9233.4125.0834.75
RMSE50.6330.71186.2569.6594.9350.1148.945.97105.65
MBIAS1.664.2823.716.6819.973.8816.77.2112.43
ECLIPSEMean20108.995.8337.3617.5331.9111.0312.7111.4421.17
RMSE28.6221.132.9626.1225.5622.6418.1620.4226.30
MBIAS−17.71−14.37−1.28−10.97−0.74−16.45−10.93−11.66−8.39
Mean20157.814.7730.4315.3126.869.7311.1410.0317.70
RMSE25.1216.128.1320.5421.0815.4912.4815.0321.13
MBIAS−14.74−10.891.65−6.322.16−11.31−5.57−7.84−4.63
HTAPMean201066.9943.3173.355372.558.3674.5549.6660.93
RMSE79.3762.0689.9656.0988.676.59106.6165.0876.65
MBIAS40.2923.1234.7124.4939.8530.8850.9126.5631.37
Mean201557.9334.1357.4941.2254.5550.2751.4837.7847.56
RMSE73.3656.1680.9648.8173.770.5289.0955.0467.46
MBIAS35.3918.4728.7119.5929.8629.2334.7719.9125.23
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Tatsumi, K.; Diep, N.T.H. Validation of Anthropogenic Emission Inventories in Japan: A WRF-Chem Comparison of PM2.5, SO2, NOx and CO Against Observations. Data 2025, 10, 151. https://doi.org/10.3390/data10090151

AMA Style

Tatsumi K, Diep NTH. Validation of Anthropogenic Emission Inventories in Japan: A WRF-Chem Comparison of PM2.5, SO2, NOx and CO Against Observations. Data. 2025; 10(9):151. https://doi.org/10.3390/data10090151

Chicago/Turabian Style

Tatsumi, Kenichi, and Nguyen Thi Hong Diep. 2025. "Validation of Anthropogenic Emission Inventories in Japan: A WRF-Chem Comparison of PM2.5, SO2, NOx and CO Against Observations" Data 10, no. 9: 151. https://doi.org/10.3390/data10090151

APA Style

Tatsumi, K., & Diep, N. T. H. (2025). Validation of Anthropogenic Emission Inventories in Japan: A WRF-Chem Comparison of PM2.5, SO2, NOx and CO Against Observations. Data, 10(9), 151. https://doi.org/10.3390/data10090151

Article Metrics

Back to TopTop