Air pollutant emissions play a determinant role in deteriorating
air quality. However, an uncertainty in emission inventories is still the key
problem for modeling air pollution. In this study, an updated emission
inventory of coal-fired power plants (UEIPP) based on online monitoring data
in Jiangsu Province of East China for the year of 2012 was implemented in the
widely used Multi-resolution Emission Inventory for China (MEIC). By
employing the Weather Research and Forecasting model with
Chemistry (WRF-Chem), two simulation experiments were executed to assess the
atmospheric environment change by using the original MEIC emission inventory
and the MEIC inventory with the UEIPP. A synthetic analysis shows that power
plant emissions of PM
East China is one of the regions with serious air pollution and frequent haze. In these highly polluted regions, air pollutant emissions play a key role in air quality, and their variations can cause a large uncertainty in air pollution modeling and prediction. It is also crucial for air pollution mitigation to comprehensively understand air pollutant emissions and their impacts on the atmospheric environment. Emission inventories are essential for atmospheric environment research, especially for modeling studies and air quality policy decisions.
During past decades, emission inventories for China were established by
several groups. These include the global-scale studies, such as the Reanalysis
of the Tropospheric chemical composition (RETRO; Schultz, 2007; Zheng et
al., 2009), the Hemispheric Transport of Air Pollution (HTAP; Janssens-Maenhout et al., 2015)
and the Emission Database for Global
Atmospheric Research (EDGAR), and the national-scale studies including the
Transport and Chemical Evolution over the Pacific mission (TRACE-P; Olivier
et al., 2005), the Intercontinental Chemical Transport Experiment-Phase
B (INTEX-B; Zhang et al., 2009), the Regional Emission inventory in
Asia (REAS; Ohara et al., 2007) and the Multi-resolution Emission Inventory
for China (MEIC,
Air pollution in East China is changing from coal-smoke to mix-source
pollution, particularly the secondary aerosols surging in severe haze
episodes (R.-J. Huang, et al., 2014), with more complicated chemical
reactions involved in the interaction of particle formation and O
Jiangsu Province is one of the most developed areas in East China, providing
a living place for a population of 79.2 million people with the highest gross
domestic production per capita in China (NBS, 2013; JSNBS, 2013). Severe air
pollution episodes of haze and photochemical pollution have shrouded this
province in recent years (Fu et al., 2008; Wang et al., 2014; Qi et al.,
2015). As elevated emission sources, power plants emit air pollutants with
longer life cycles in the upper air and more efficiently transport across
regions because of less deposition driven by stronger winds and a well
organized circulation in the upper air, e.g., by low-level jets (Hu et al.,
2013). This leads to more significant environmental effects than surface
emissions (e.g., on-road emissions), reflecting the potential importance of
accurately estimating the power plant emissions and their influences on air
quality. Based on a unit-based methodology, Zhao et al. (2008) developed an
inventory of coal-consuming power plants for all the provinces in China. The
annual SO
Y. Zhang et al. (2015) established an emission inventory of coal-fired power plants (UEIPP) by collecting the online monitoring data from power plants in atmospheric verifiable accounting tables of Jiangsu Province for 2012. The volumes of flue-gas and pollutant concentrations were measured on-site for each unit, providing the more realistic data for calculating power plant emissions than those used in previous studies. As a major objective of this study, the UEIPP was integrated into MEIC to evaluate the impact of updated emissions on the regional atmospheric environment with an air-quality model. We present the details of model settings, observational data and emission inventories in Sect. 2 and the modeling evaluation in Sect. 3. The impact of emissions change on the atmospheric environment and the underlying mechanism are discussed in Sect. 4. The study is summarized in Sect. 5.
Model domain and the locations of the 13 cities in Jiangsu Province.
The period from 29 November to 31 December 2013 was chosen as the modeling period, covering a severe haze period (from 3 to 8 December 2013) in Jiangsu Province. The online coupled Weather Research and Forecasting Model with Chemistry (WRF-Chem in version 3.7.1) was configured in three nesting domains with horizontal resolutions of 45 km covering most areas of east Asia, 15 km covering eastern China and surrounding areas, and 5 km covering Jiangsu Province and surrounding areas (Fig. 1a). Vertically, there were 35 full eta levels from surface up to 100 hPa with 7 levels below 1 km. The National Center for Environmental Prediction Final Global Forecast System operational analysis data (Kalnay et al., 1996) was utilized for providing the initial and lateral meteorological conditions to WRF-Chem. Grid nudging (Stauffer and Seaman, 1990) was employed for the outermost domain every 6 h (treating temperature, horizontal wind, and water vapor) to guarantee the precision of large-scale meteorology during the simulations.
The selected physical configurations included the Morrison double-moment
microphysics scheme (Morrison et al., 2009), RRTMG (Rapid Radiative Transfer
Model for GCMs – global climate models), long- and short-wave radiation scheme
(Iacono et al., 2008), Grell 3-D cumulus parameterization, the Yonsei University
planetary boundary layer scheme (Hong et al., 2006), and the Noah land surface
model. For the chemistry and aerosol mechanisms, the CBM-Z (Carbon Bond Mechanism version Z;
Zaveri, 1999) coupling with the 8-bin sectional MOSAIC (Model for Simulating
Aerosol Interactions and Chemistry) with aqueous chemistry (Zaveri et
al., 2008) was used. The MOSAIC, treating all the important aerosol
components including nitrate, sulfate, ammonium, black carbon (BC), and primary
organic aerosols and other inorganic aerosols, is efficient without
compromising accuracy and widely used in air-quality and regional or global
aerosol models (Zaveri et al., 2008). Since the MOSAIC is incapable of
simulating secondary organic aerosols (SOAs), the simulated organic aerosols
mentioned hereinafter all refers to primary organic aerosols. The crucial
processes of radiation feedback, aerosol and cloud interaction, dry
deposition, wet scavenging and cloud chemistry were turned on. Biogenic
emissions were calculated online with MEGAN (Model of Emission of Gases
and Aerosol from Nature; Guenther et al., 2006). The initial and boundary
chemistry conditions were based on the vertical profiles of O
Meteorological fields simulated by the model are crucial for the accuracy of
air-quality modeling. In the southern, middle, and northern parts of Jiangsu
Province, we selected the three prefecture-level cities of Nanjing, Yancheng, and
Lianyungang, respectively, to evaluate the overall perspective of
meteorological simulation with the available observations. The observed
meteorological data, involving 2 m temperature, 2 m relative humidity and
10 m wind speed and direction were collected from the Jiangsu Provincial
Meteorological Bureau and Meteorological Information Comprehensive Analysis
and Process System (MICAPS) of the China Meteorological Administration (CMA).
Hourly surface concentrations of chemical species in 13 cities of Jiangsu,
including SO
This study uses the MEIC inventory as the default anthropogenic emissions
including the emissions of sulfur dioxide (SO
The UEIPP in Jiangsu Province for the year of 2012 consisting of six online
species (SO
The SO
The annual emissions of CO and NMVOCs were calculated using Eq. (3):
Following the method used by Li et al. (2014) and the mechanism-dependent
mapping tables developed by Carter (2013), the NMVOCs in UEIPP were specified
to individual constituents in the Regional Acid Deposition Model chemical
mechanism (RADM2; Stockwell et al., 1990), which could be adapted to the
WRF-Chem/CBM-Z mechanism used in this study. The primary distinction between
UEIPP and the power plant emission inventory estimated in previous China
studies lies in the different data used and subsequently the estimation
algorithm as well as the temporal resolution. Previously, the power plant
emission inventory was mostly estimated using various data such as activity
levels, boiler types, fuel types, control policies, and emission factors, and
the activity levels were usually collected on annual or monthly timescales.
In the UEIPP, the emissions of SO
The UEIPP and MEIC power plant emissions of major air pollutants as well as the ratios in total emission inventory over Jiangsu Province in 2012.
Statistics between observed (Obs.) and modeled (Mod.) meteorology.
The UEIPP and MEIC power plant emissions of major air pollutants with their
fractions in the total emissions over Jiangsu Province in 2012 are presented
in Table 1. Appreciable differences between the two
power plant emission inventories were revealed. The power plant emissions of
SO
Statistics comparing observed and simulated PM
Differences in power plant emissions between MEIC and UEIPP in 2012 (units: tons).
The spatial difference of the two emission inventories over Jiangsu Province is
shown in Fig. 2, as well as their absolute values in Fig. S1 in the Supplement. The UEIPP
presented low emissions of SO
To assess the simulation performance with the UEIPP and changes in the atmospheric environment over Jiangsu Province under the updated emission conditions, simulations with the original MEIC emission inventory (hereinafter referred to as the MOD1 simulation) and the updated MEIC emission inventory with the power generation replaced by UEIPP (hereinafter referred to as the MOD2 simulation) were carried out. The difference in chemical components between MOD1 and MOD2 simulations were used to assess atmospheric environmental changes in the following sections.
An evaluation of the meteorological simulations over the domain with 5 km
horizontal resolution was carried out in regards to temperature, relative
humidity (RH), and wind speed and direction in Nanjing, Yancheng, and
Lianyungang. The evaluation of statistical parameters included mean
bias (MB), correlation coefficient (
The spatial distributions of near-surface SO
Validation statistics of SIAs in PM
Differences in chemical species between MOD2 and MOD1 in December 2013.
Unit: “pptv” for OH; “ppmv” for VOC; “
Chemical species of PM
The surface observations of PM
The
Spatially, overestimates of SO
As mentioned above, MOD2 with the introduction of UEIPP improved the
simulation of air pollutants, especially of O
Aside from introducing the updated emission inventories, another important
and meaningful work in this study is to explore how the emission changes
affect the atmospheric environment, especially in severe haze episodes, for
understanding the complexity of the atmospheric environment. To this end, we
presented the differences in atmospheric compositions simulated with MOD1 and
MOD2 in Fig. 4. Consistently with the emission changes (Fig. 2), the
concentrations were reduced for SO
Increased concentrations (
Ratios of HCHO
As a secondary air pollutant in the boundary layer, O
Daily variations in
Daily variations in
Diurnal variations in
Quite surprising to us, the surface PM
As described in Sect. 4.1, the declined emissions of primary PM
As major oxidizing agents in the atmosphere, O
To evaluate how the formation of secondary aerosols responded to the enhanced
oxidizing capacity, we analyzed the BC-scaled concentrations for sulfate and
nitrate. The purpose was to eliminate the influence of air pollutant dilution
and mixing in the atmospheric physical process. Since BC is quite
inert to chemical reactions, its variations could well reflect the atmospheric
physical processes. Thus, the BC-scaled concentration could better represent
the contribution of chemical reactions (Zheng et al., 2015). Figure 7 presents
the daily averaged variation in BC-scaled concentrations for sulfate and
nitrate, appending the differences in O
The SOR (molar ratio of sulfate to sum of sulfate and SO
As shown in Table 5, the total concentration of sulfate, nitrate and ammonium
increased by 1.32
It should interpret the larger enhancement in concentrations of SIAs than
the PM
Power plants, as major air pollutant sources in China, have been imposed with restrictions by the government in response to the increasing air pollution, which led power plant emissions to large variations during the past few years. Due to various underlying data and approaches, there remained uncertainties in estimating the power plant emissions. In the present study, the UEIPP in Jiangsu Province for 2012 was introduced into the MEIC emission inventory as the major point sources. The variation and complexity of the atmospheric environment in response to the change of power plant emissions over Jiangsu were studied by executing the WRF-Chem simulations using the original emissions of MEIC and the MEIC with its power plant emission inventory updated by the UEIPP.
This study focused on the uncertainties in estimating the power plant
emissions. In the UEIPP, the emission amounts of SO
The UEIPP drove the simulation performance to be superior to the original power
plant emission of the MEIC inventory in terms of the proximity between simulated
and observed air pollutant concentrations, suggesting a more realistic power
plant emission inventory. The complexity of the atmospheric environment was
also revealed by comparing the changes in various primary and
secondary compositions in the atmosphere. Atmospheric oxidizing capacity was
reinforced in response to the enhancement of O
Given the complicated processes in environmental change, the restrictions of emissions should be comprehensively considered rather than one single factor. Furthermore, the effects of emission inventories on air-quality variations could be assessed based on long-term observation and simulation studies, and formation of SOAs would also enhance due to the reinforced atmospheric oxidizing capacity and higher VOC emissions.
Data used in this manuscript can be provided upon request by email to Lei Zhang (leiz7002@126.com) or Tianliang Zhao (josef_zhao@126.com).
The authors declare that they have no conflict of interest.
This study was jointly funded by National Key R & D Program Pilot Projects of China (2016YFC0203304); Key program for Environmental Protection of Jiangsu Province in 2015 (2015017); Program of Multi-Scale Haze-fog Modeling Development of National Science & Technology Support Program (2014BAC16B03); and the Program for Postgraduates Research and Innovation in Universities of Jiangsu Province (KYLX16_0937). Edited by: Renyi Zhang Reviewed by: two anonymous referees