Inverse estimation of NO_x emissions over China and India 2005–2016: contrasting recent trends and future perspectives

Check for updates using Crossmark

Buy this article in print

We developed an inversion system (Yumimoto et al 2015) for NO_x by extending the method proposed by Martin et al (2006) and Lamsal et al (2011). Based on the linear relationships between NO_x emissions (E) and NO₂ columns (Ω), Martin et al (2006) estimated NO_x emissions with coefficient α taken from the model simulation by using the relationship

For example, our previous study also used the linear relationships between NO_x and NO₂, and we estimated NO_x emissions from China (Itahashi et al 2014). In the inversion system developed in the present study, we assumed that the linear relationships between NO_x and NO₂ can be obtained over the specific range of emissions variation from E₀; hence, we extended equation (1) to

Coefficient β, which links the NO_x emissions and NO₂ column density, was obtained via the Community Multi-scale Air Quality (CMAQ) sensitivity simulation. We evaluated this coefficient by setting the increase in NO_x emissions as 20% in the sensitivity simulation.

The NO_x emissions and NO₂ column in the base-case simulation are represented by E_f and Ω_f, and those in the sensitivity simulation are represented as ${E}_{f}^{{\prime} }=(1+\delta ){E}_{f}$ and ${{\rm{\Omega }}}_{f}^{{\prime} },$ respectively. Thus, equation (2) can be rewritten as

and β can be expressed as

where δ is 0.2 (20%). Then, the NO₂ column was estimated based on the linear unbiased optimum estimation by assimilating the observed NO₂ column. We used OMI observations from the Dutch OMI tropospheric NO₂ (DOMINO) data product version 2.0 (Boersma et al 2011) to constrain the NO₂ column. Level 2 swath data from the DOMINO product was applied with limits of clear sky conditions and good data-quality flags. The observed and modeled NO₂ column in the base-case simulation are Ω_o and Ω_f, respectively. The errors were assumed to follow a lognormal distribution. Error ε₀ includes errors from the measurements and model, and ε_f is the background error. Therefore, the a posteriori column density (Ω_a) and error (ε_a) can be estimated as follows

We used an additional treatment for Ω_f of the modeled NO₂ column in the base-case simulation. Our previous study (Yumimoto et al 2015) found that the biases caused by the errors in the modeling system can introduce artificial biases in the inversely estimated emissions. To reduce the effect of the model bias, we used ${\hat{{\rm{\Omega }}}}_{f}={{\rm{\Omega }}}_{f}-B$ instead of Ω_f in equations (3)–(5), where B is the estimated model bias. Finally, based on the analogous relation between the a priori (Ω_f) and a posteriori (Ω_a) estimated column densities in equation (3), a posteriori emissions (E_a) corresponding to a posteriori Ω_a were calculated

As the a priori emissions (E_f), we used the Regional Emission inventory in ASia (REAS) version 2.1 (Kurokawa et al 2013), which includes anthropogenic emissions (e.g. NO_x, SO₂, and CO etc) over Asia. Because REAS covers the years 2000–2008, the emissions in 2008 were used as a priori emissions after 2008 (i.e. 2009–2016). To estimate the long-term emissions from 2005 to 2016, the monthly mean NO₂ column was used and optimized in each grid. The model biases were derived from the 4 year comparison between the NO₂ columns modeled with the a priori emissions and the observed NO₂ columns during 2005–2008 estimated on a monthly timescale at each grid.

The grid to estimate the a posteriori emissions is based on the grid where the NO_x emissions are larger than 1 ×ばつ 10¹⁵ molecules cm⁻² s⁻¹, anthropogenic emissions are attributed as larger than 90%, and observations were conducted on 50 d per year. For background error ε_f, the uncertainty provided as percentages in REAS version 2.1 emissions (table 13 of Kurokawa et al 2013) was multiplied by ${\hat{{\rm{\Omega }}}}_{f}.$ For observation error, we used the information in the observation dataset (retrieval uncertainty) and the representation errors estimated from the standard deviation of the grid averaging. As shown in the following section, Chinese anthropogenic emissions rapidly increased from 2009 to 2011, reflecting economic development. Using the REAS emissions in 2008 as a priori emissions in this period may cause negative bias in the a posteriori emissions because the a priori emissions are much lower than the actual emissions. To overcome this limitation, for Chinese emissions from 2009 to 2011, we developed the sequential update technique, in which the a posteriori emissions in the previous year are set as the a priori emissions (i.e. perform the inverse modeling in 2010 with the a posteriori emissions in 2009 as the a priori emissions).

Our inversion system has four new features. First, the method is simple for updating and extending the bottom-up emissions quickly with relatively light computational burden. Second, the errors of the model and observations are included in the inverse estimation by applying a linear unbiased optimum estimation. Third, the model bias is considered by introducing it into the modeled NO₂ column. Fourth, the sequential update technique captures the rapid increase of Chinese emissions. The consideration of the model bias by introducing B and the development of the sequential update technique are an extension of previous works by Martin et al (2006) and Lamsal et al (2011).

Our previous study (Yumimoto et al 2015) partly applied this inverse modeling system to extend the bottom-up emissions and to investigate the long-term trend in China during 2009–2012. The tropospheric NO₂ columns predicted with the estimated emissions reproduced the spatial distribution, seasonal cycle, and inter-annual variation in the observations, and achieved better agreement with the satellite-observed columns than those predicted with the bottom-up emissions.

The forward model simulation was performed with the regional chemical transport model simulation of the CMAQ modeling system version 4.7.1 (US EPA Office of Research and Development 2010). To drive CMAQ, meteorological fields were prepared with the Weather Research and Forecasting (WRF) model version 3.3 (Skamarock et al 2008). Analysis nudging was conducted to calculate the WRF model by using the Final Analysis Dataset of National Centers for Environmental Prediction (NCEP FNL) available with 1° ×ばつ 1° horizontal resolution with 6 h intervals. The model domain covers the whole of East Asia with a 60 km horizontal grid resolution with 135 ×ばつ 135 grids centered at 30°N and 110°E on a Lambert conformal projection. The vertical grid on sigma-pressure coordinates was extended to 100 hPa with 34 layers by nonuniform spacing. Each year, the simulation was conducted with a 15 d model spin-up run. Emissions were set as follows. Anthropogenic emissions over Asia were obtained from REAS (see section 2.1) and those over Japan were obtained from the Japan Auto-Oil Program (JATOP 2012a, 2012b). Biogenic emissions were prepared from the Model of Emissions of Gases and Aerosols from Nature (Guenther et al 2012), and biomass burning emissions were obtained from the Global Fire Emission Database (van der Werf et al 2006). Volcanic activity data were taken from the Aerosol Comparisons between Observations and Models (Diehl et al 2012) and the Japan Meteorological Agency (Kazahaya et al 2001). Lateral boundary conditions were obtained from the monthly mean of the global chemical transport model called CHASER (Sudo et al 2002). This modeling system was used to study the air quality in Asia in our previous work (e.g. Morino et al 2015, 2017). In addition to these base-case simulations, sensitivity simulations were conducted by increasing all NO_x emissions sources by 20% to obtain the sensitivity coefficients (β in equation (4)), which combine the NO₂ column response according to NO_x emissions change. These sensitivity simulations were also conducted from 2005 to 2016.

Temporal variation of the OMI NO₂ column and inverse estimation of NO_x emissions over China with the estimated error are shown in figure 1. The emissions from the updated REAS inventory (Kurokawa et al 2018), MEIC inventory (Li et al 2017b, Zheng et al 2018), and inverse estimation by Miyazaki et al (2017) are also shown. In 2005, NO_x emissions were around 20 Tg and other estimations were within the range of 15–20 Tg (Reis et al 2009). All inventories showed increasing trends that peaked around 2011, and then decreased. The NO₂ column observation and emissions decreased from 2008 to 2009 owing to the economic recession (e.g. Lin and McElroy 2011, Itahashi et al 2014). After the ecomomic recovery, the observed NO₂ column and NO_x emissions increased again until 2011. Our inverse estimation corresponded well with the other emission estimates, and reached its peak of 29.5 Tg yr⁻¹ in 2011. These emission variations before and after 2011 are summarized in table 1. NO_x emissions increased by 1.0–1.7 Tg yr⁻¹ (3.9%–7.1%/year) from 2005 to 2011, and decreased by around 1.0 Tg yr⁻¹ (around 3%/year) after 2011. The NO_x emissions in 2016 generally corresponded to around 2008 levels. These reduction trends over China were due to the widespread use of denitrification units, low-NO_x burners in large new power plants, and new vehicle emission standards (Kurokawa et al 2018, Zheng et al 2018).

The variation over China is shown in figure 1, and we focused on eastern China, particularly over the centers of population and economic activity. Figure 2 shows the annual trends from 2011 to 2016. Generally, there were decreasing trends, although there were some increasing trends over parts of Sichuan Province, Fujian Province, and northeast China. The observed and modeled NO₂ column showed similar variations and the trends with the largest changes of greater than −10%/year were centered on eastern China (i.e. Henan, Hubei, and Anhui Provinces). The inverse estimated emissions on the province scale showed trends of around −3%/year and there were large differences according to location. Our inverse estimate of NO_x emissions from China as a whole showed the peak in 2011, whereas the estimation by Zheng et al (2018) showed the peak in 2012 (figure 1). Figure 3 shows the peak year of the observed and modeled NO₂ column and inverse estimated NO_x emissions on the province scale in China. The peak year of the observed and modeled NO₂ column corresponded well; generally, the peak year was around 2011 (light orange), Beijing and Guangzhou exhibited an earlier peak between 2005 and 2008 (blue to green), and northeast China exhibited a later peak between 2013 and 2016 (red).

The observed NO₂ column and estimated NO_x emissions over India with estimated error are shown in figure 4. NO_x emissions increased continuously over India from 2005 to 2016. In contrast to China, NO_x emissions increased in most parts of India (results not shown). Other emissions taken from the updated REAS inventory (Kurokawa et al 2018), the inverse estimation by Miyazaki et al (2017), the Emission Database for Global Atmospheric Research (EDGAR) version 4.3.1 (http://edgar.jrc.ec.europa.eu/index.php), and the emissions dataset from a technology-linked inventory over India (Pandey et al 2014, Sadavarte and Venkataraman 2014) are also shown. The analyzed periods were 2005–2010, 2005–2015, and 2005–2016, and the variations in our estimation and other emission estimations are summarized in table 2. The changes were +0.1 to +0.6 Tg yr⁻¹ (+1 to +6%/year) from 2005 to 2010 and +0.2 to +0.6 Tg yr⁻¹ (+2 to +6%/year) from 2005 to 2015 or 2016. Our estimate shows a different trend from that in the observed NO₂ column during 2014–2015. One possible reason for this is meteorological variability. The meteorological impacts on the relationship between NO_x emission and tropospheric NO₂ column amount should be analyzed in detail in future studies. In our estimation, there was a large increase during 2013–2014. The latest EDGAR for greenhouse gases emission inventory (https://edgar.jrc.ec.europa.eu/overview.php?v=CO2andGHG1970-2016) also report a similar large increase in CO₂ emissions (ca. 10%/year) during 2013–2014, and this year-on-year growth rate is greater than in other years (ca. 5%–6%) between 2005 and 2016. Therefore, the increase in fossil fuel consumption could be one plausible reason for the large increase of our estimated NO_x emissions during 2013–2014. Although there were differences in the rate of increase, all estimations indicated a continuous increase in emissions over India, in contrast to the varying trend over China.

Our inverse NO_x estimation over China and India showed a reversal in the increase in emissions after 2011 over China and a continuous increase over India. In this section, we provide perspectives for future emission changes in China and India. Based on the inverse estimation in this study, the NO_x emission trend was −0.83 ± 0.14 Tg yr⁻¹ over China and +0.76 ± 0.16 Tg yr⁻¹ over India with a significance level of p < 0.005 (Students’ t-test) during 2011–2016 (figure 5). The peak year of NO_x emission from China was set as 2011 in the trend analysis. Future scenarios reported in several other studies are also included in figure 5. Chatani et al (2014) (shown in green in figure 5) estimated three scenarios based on business as usual (BAU), additional legislation and technological developments (PC0), and legislation and developments beyond those in PC0 (PC1) over China and India in 2030 based on 2010 emissions. The European Union’s project of Evaluating the Climate and Air Quality Impacts of Short-Lived Pollutants (shown in yellow in figure 5; Stohl et al 2015) created current legislation (CLE) and short-lived climate pollutant mitigation (MIT) scenarios from 2015 to 2050 over China and India. Venkataraman et al (2018) (shown in red in figure 5) reported emission pathways based on reference (REF), aspirational (S2), and ambitious (S3) scenarios from 2015 to 2050 over India.

Over China, current reduction trends after 2011 are beyond the expected regulation. Both the PC0 and MIT scenarios foresaw almost flat trends for NO_x emissions from China; however, all top-down and bottom-up estimates after 2011 suggested declining trends (figure 1 and table 1). The current decreasing trends are consistent with the projected regulation in the PC1 scenario with decreases of 0.7 Tg yr⁻¹ (3.9%/year). In contrast, the increasing trends over India match the unregulated scenarios of 0.7 Tg yr⁻¹ (5.5%/year) for BAU, 0.3 Tg yr⁻¹ (2.9%/year) for CLE, and 0.6 Tg yr⁻¹ (4.2%/year) for REF. Based on the confirmation of our recent estimation trends over China and India with various future scenarios, we assume continuous trends of −0.83 ± 0.14 Tg yr⁻¹ over China and +0.76 ± 0.16 Tg yr⁻¹ over India after 2016. Under this assumption, NO_x emissions from China and India will be similar in 2023 with around 19–20 Tg yr⁻¹, and NO_x emissions from India will surpass those from China in 2024. Based on the range of the trends, the excess NO_x emissions from India rather than China will occur between 2022 (maximum decrease of −0.97 Tg yr⁻¹ over China and maximum increase of +0.92 Tg yr⁻¹ over India) and 2025 (minimum decrease of −0.69 Tg yr⁻¹ over China and minimum increase of +0.60 Tg yr⁻¹ over India). Because the US, which is the second largest source, has shown a decline of −0.65 Tg yr⁻¹ during 2011–2016 (https://epa.gov/air-emissions-inventories/air-pollutant-emissions-trends-data) or even in the possible slowdown trend (Jiang et al 2018), India will become the world’s largest NO_x emission source. These are the first perspectives for future emission changes to two importance sources in Asia, and we should continue to monitor these changes closely.