Simultaneous Dust and Pollutant Transport over East Asia: The Tripartite Environment Ministers Meeting March 2014 Case Study

Transcription

1 SOLA, 2017, Vol. 13, 47 52, doi: /sola Simultaneous Dust and Pollutant Transport over East Asia: The Tripartite Environment Ministers Meeting March 2014 Case Study Itsushi Uno 1, Keiya Yumimoto 2, Xiaole Pan 1, 6, Zhe Wang 1, 6, Kazuo Osada 3, Syuichi Itahashi 4, and Shigekazu Yamamoto 5 1 Research Institute for Applied Mechanics, Kyushu University, Kasuga, Fukuoka, Japan 2 Meteorological Research Institute, Tsukuba, Ibaraki, Japan 3 Graduate School of Environmental Studies, Nagoya University, Nagoya, Japan 4 Central Research Institute of Electric Power Industry, Abiko, Chiba, Japan 5 Fukuoka Institute of Health and Environmental Sciences, Dazaifu, Fukuoka, Japan 6 Institute of Atmospheric Physics, CAS, Beijing, China Abstract Heavy dust storms that occurred between 13 and 23 March 2014 were selected for analysis as a case study of dust and sand storm events by the Tripartite Environment Ministers Meeting. During this period, two dust and anthropogenic aerosol events were observed and analyzed. The GEOS-Chem chemical transport model, which includes dust-acid uptake processes, successfully reproduced the aerosol variations and explained the Asian-scale dust-pollutant transport processes. Our results confirmed the importance of coarse-mode dust-nitrate as evidence of polluted dust pollution. The model analysis showed that the formation of dust-nitrate occurred over the Yellow Sea and East China Sea before arriving in Japan. We showed that more than 40% of nitrate exists in dust-nitrate when air mass arrived in Japan. (Citation: Uno, I., K. Yumimoto, X. L. Pan, Z. Wang, K. Osada, S. Itahashi, and S. Yamamoto, 2017: Simultaneous dust and pollutant transport over East Asia: The tripartite environment ministers meeting March 2014 case study. SOLA, 13, 47 52, doi: /sola ) 1. Introduction Heavy anthropogenic pollution and natural mineral dust events cause significant environmental problems in East Asia (e.g., Huang et al. 2014; Zhang et al. 2003). The dust source regions in Asia are located upwind of major anthropogenic pollutant emission regions. Therefore, during transport from dust source region to Japan, dust is easily mixed with air pollutants, and reacts heterogeneously producing dust-sulfate and dust-nitrate. These aerosols mainly exist in the coarse mode (e.g., Song and Carmichael 2001). Such a mixture of dust and pollution is referred to as polluted dust (Mori et al. 2003), and causes significant health problems (e.g., Kanatani et al. 2010). Because NOx emissions in East Asia have rapidly increased over the last decade, our understanding of the formation of polluted dust has become increasingly important. The continuous measurement of the coarse-mode aerosol composition is therefore important, but has not been sufficiently reported in the literature. A heavy dust storm occurred between 13 and 23 March 2014 and was selected as a case study of a dust and sand storm event by working group-1 of the Tripartite Environment Ministers Meeting (TEMM). During this period, two dust events were observed in Seoul and Fukuoka, Japan. In both cases, we observed high concentrations of dust and anthropogenic aerosols. In this study, we analyzed 1-h interval dust and anthropogenic aerosol (both fine- Corresponding author: Itsushi Uno, Research Institute for Applied Mechanics, Kyushu University, Kasuga, Fukuoka , Japan. uno@riam.kyushu-u.ac.jp. 2017, the Meteorological Society of Japan. and coarse-mode) observation results and numerically simulated the Asian-scale transport process during the case study period. 2. Data and study method 2.1 Observation data Mass concentrations of particulate matter (PM), sulfate and nitrate in both the fine (Dp < 2.5 μm) and coarse mode (2.5 μm < Dp < 10 μm) were measured using a continuous dichotomous aerosol chemical speciation analyzer (ACSA-12, Kimoto Electric Co., Ltd) at 1-h time intervals on the rooftop (4 th floor) of the Fukuoka Institute of Health and Environmental Science (longitude: E, latitude: 33.5 N). The details of the ACSA-12 instrument have been published elsewhere (Kimoto et al. 2013). In addition to the ACSA-12 measurement, fine-mode aerosols (sampled by FRM-2025i, ThermoScietific) were also collected with a polytetrafluoroethylene filter (FP) at Fukuoka, with a sampling interval of 6-h. The samples were analyzed by ion chromatography (IC) to determine the concentration of NO 3, SO 4 2, NH 4 +, water soluble Ca 2+ and Na + (FP/IC). From the ground-based PM 2.5 and PM 10 observations shared in TEMM, hourly PM 2.5 and PM 10 concentrations (using beta-ray absorption method) measured in Seoul, Korea were used for the validation of model results. Asian Dust and aerosol lidar observation Networks (AD-Net) (Sugimoto et al. 2008) measure both anthropogenic and natural aerosols continuously. This network provides vertical profiles of aerosol backscattering at 532/1064 nm and the depolarization ratio at 532 nm with high spatial and temporal resolution. The automated AD-Net lidar retrieval data for the total aerosol extinction coefficient (Shimizu et al. 2010) at Fukuoka and Seoul were compared to the modeled aerosol variations. The atmospheric aerosol particles (0.5 μm < Dp < 10 μm) were measured using a polarization optical particle counter (POPC) (YGK Corp., Yamanashi, Japan) at the Kyushu University, Fukuoka, Japan. The POPC uses a 780-nm linearly polarized laser. The size of particles is determined from the forward scattering intensity; two polarized compounds (s-polarized/p-polarized: polarization direction perpendicular and/or parallel to the plane of the scattering angle, respectively) are measured simultaneously from the backward scattering. The depolarization ratio (DR, the fraction of s-polarized to the total backward scattering, S/(S+P)) is a good indicator of particle sphericity. The overall measurement uncertainty for the number density of super micron particles was less than 15%. Details of the POPC have been reported by Kobayashi et al. (2014). 2.2 Chemical transport model analysis We used the GEOS-Chem 3-D chemical transport model (version 09-02) (Bey et al. 2001; Park et al. 2004) to simulate the aerosols, including mineral dust and secondary inorganic aerosol formation over East Asia, which were nested into a global model. For this study, the reactive uptake of HNO 3 and SO 2 on dust,

2 48 Uno et al., Simultaneous Dust and Pollutant Transport over East Asia which was limited by dust alkalinity, and the uptake of gas-phase H2SO4, which was limited by competition with other aerosol surfaces (Fairlie et al. 2010) were used. In this version, dust-nitrate (typically as Ca(NO3)2) was simulated for each dust bin separately based on the heterogeneous reaction between dust and nitric acid gas, as follows: CaCO3 + 2HNO3 Ca(NO3)2 + H2O + CO2 The reactive uptake constant is a function of RH, dust diameter, and HNO3 gas concentration (Fairlie et al. 2010). A similar heterogeneous reaction with sea salt was also used. Supplement-1 shows detailed formulation. Hereafter we denoted regular nitrate (i.e., NH4NO3) as A-NO3 and dust-nitrate as D-NO3, and sum of them simply as NO3. Dust in GEOS-Chem is distributed in four size bins (radii , , , and μm). Sea salt is distributed in two size bins (dry radii and µm). An assimilated dust emission factor and several modifications to the dust emission fraction within dust-bins, and the change of wet scavenging efficiency proposed by Yumimoto et al. (2017) were used in this study. In the analysis of the PM concentration, PM2.5 and PM10 were calculated by the summation of the individual aerosol and dust components of the model. The model uses the assimilated meteorological fields from the Goddard Earth Observing System 5 (GEOS5) of the United States National Aeronautics and Space Administration (NASA). The model has a horizontal resolution of for a global run and for Asian one-way nesting runs (70 E 150 E, 11 S 55 N), both with 47 vertical levels from the surface to 0.01 hpa. The lowest model layer thickness was about 130 m. We used anthropogenic emissions from the Emission Database for Global Atmospheric Research (EDGAR) for the global domain and the Regional Emission inventory in ASia (REAS) ver. 2.1 for the Asia domain, as estimated by Kurokawa et al. (2013). The global simulation started from the January 2013 to March 2014, and Asian simulation using global concentrations started from January The other basic settings were the same as in Uno et al. (2013, 2016). 3. Results and discussion 3.1 Horizontal distribution and PM2.5 and PM10 variation Surface Synoptic Observations (SYNOP) reported two dust events in inner-mongolia in China and Mongolia between 11 and 17 March Figure 1 shows the daily change in the horizontal distribution of dust and SO42 from March 12 to 19. The symbol $ indicates that a dust phenomenon was observed at the SYNOP site. Figure 2 shows the (a) Time-height (TH) plot of the lidar total Fig. 1. Daily changes in dust (yellow) and SO42 (contours and pink shade) at 0900JST. $ indicates the Surface Synoptic Observations dust reporting site. Two back-trajectories from Fukuoka for periods A1 and A2 are also overlayed. Open circled trajectory in (d) (f ) is upper path and close circled trajectory in (d) (f ) is lower path.

3 SOLA, 2017, Vol. 13, 47 52, doi: /sola extinction coefficient profile at Fukuoka, (b) the time variation of PM 2.5 by ACSA-12, GEOS-Chem and lidar total extinction coefficient (averaged between m), (c) PM 10 by ACSA- 12, GEOS-Chem, and the POPC particle volume concentration in the µm size range. Figure 3 shows the same information as Figs. 2b and 2c, but for Seoul. We observed two dust episodes in Fukuoka and Seoul. We labeled the two episodes A1 and A2 as shown in the figures. The location of the daily position of each dust cloud is shown in Fig. 1 (as A1 and A2). The first dust episode A1 (emitted on March from the southern part of Mongolia and Inner Mongolia of China) moved slowly and was trapped within a high pressure system. The time variation of the dust concentration in Seoul and Fukuoka shows a similar wide bell-shape, with a peak of PM 10 = 100 µg m 3 between 16 and 17 March. The line in Fig. 1c shows the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model back-trajectory starting from Fukuoka (at a 1000-m height) when the dust concentration was at a maximum. The Fukuoka area was located within the warm sector of a low-pressure system from 17 to 18 March; a rapid reduction in the dust concentration was observed due to a clean maritime air mass that penetrated from the Pacific Ocean. In contrast, the A2 dust episode (emitted on March from central Mongolia and inner Mongolia of China due to a strong outbreak of cold air) resulted in a sharp increase in the dust concentration (typically in Seoul), which was related to the quick onset of a dust cloud behind a typical cold front (Fig. 1). The peak concentration of A2 in Seoul was higher than that of Fukuoka because Seoul was located near the main dust body (A2 dust in Fukuoka was already horizontally stretched with the cold front). The line in Figs. 1d, 1e, and 1f shows the HYSPLIT backtrajectory starting from Fukuoka (at and 2000-m heights) when the dust was at a maximum. The lower air mass originated from central China (whereas the SO 4 2 was present at a high concentration ahead of the dust cloud, as shown in the pink shade), while the upper air mass was directly transported from the dust source region. We found that the modeled PM 10 showed good agreement with observations both in Fukuoka and Seoul. This shows that dust and pollution were mixed over China in A1 case, whereas in A2 case, pollutants were mainly mixed with dust over the Ocean after departing from China land. The PM 2.5 variation in Fukuoka had two peaks, which the model reproduced quite well (although it was slightly underestimated). The lidar TH plot showed that the dust layer in the A1 period was aloft (center height was around m), and the dust in the A2 period was well mixed vertically, so the agreement of ACSA-12 PM 2.5 and lidar was better for the A2 case. For PM 2.5 in Seoul, we found more than three peaks and the time variation of the observed PM 2.5 and lidar extinction showed good agreement, whereas the modeled PM 2.5 missed a couple of peaks; this indicates that the contribution from Seoul local emissions cannot be simulated completely well with the model. 3.2 Aerosol compositions in Fukuoka Figure 4 shows the aerosol chemical composition at Fukuoka. The FP/IC and ACSA-12 SO 4 2 levels were in good agreement and there were two peaks in each dust period of A1 and A2 (Fig. 4a). The maximum concentration in A2 was about 20 µg m 3. GEOS- Chem SO 4 2 levels were in good agreement with observations for the A2 period, but the model underestimated the A1 period. As shown in Fig. 1c and 1d, the footprint of air mass trajectories were different for the A1 and A2 cases. The model underestimation for A1 may be due to the underestimation of anthropogenic emission, the accuracy of meteorological parameters, and/or a possible missing/unknown chemical mechanism. Figure 4b shows the fine NO 3 levels indicated by ACSA and FP/IC. GEOS-Chem results for NO 3, A-NO 3 (red lines) and fine D-NO 3 (green line) are also shown in the figure. The total fine Fig. 2. Time series plot of (a) TH-plot of lidar aerosol total extinction (slanted area indicates the missing observation), (b) PM 2.5 and GEOS- Chem fine dust and lidar aerosol total extinction coefficient, (c) PM 10, GEOS-Chem DUSTPM 10, and a polarization optical particle counter (POPC) observation for Fuku oka. A1 and A2 indicate the two dust event periods. Fig. 3. Time series plot of (a) PM 2.5 and GEOS-Chem fine dust and lidar aerosol total extinction coefficient and (b) PM 10 and GEOS-Chem DUSTPM 10 for Seoul (PM 10 by Korean Meteorological Agency and PM 2.5 by National Institute of Environmental Research, Korea). Open stars in (a) indicate when the model under-estimated the observed peak concentration.

4 50 Uno et al., Simultaneous Dust and Pollutant Transport over East Asia Fig. 4. Time series plot of observed and modeled (a) fine SO 4 2 and RH, (b) fine NO 3, and (c) coarse NO 3, fine Na +, and fine Ca 2+ at Fukuoka. NO 3 levels indicated by GEOS Chem (red straight line) were well reproduced in A2, but were overestimated in A1. The contribution of fine D-NO 3 was smaller than the fine A-NO 3, except for the peak on 19 March. Figure 4c shows the coarse NO 3 by ACSA-12, fine Na + and Ca 2+ by FP/IC analysis; GEOS-Chem coarse D-NO 3 and sea-salt NO 3 are also shown. The model calculated total coarse NO 3 was similar during the A1 and A2 periods, but was slightly underestimated in the later part of the A2 period (this may be derived from an underestimation of the coarse dust concentration). During the A2 period, the Ca 2+ concentration was high, which indicates that Ca(NO 3 ) 2 was a major fraction of the NO 3, and the modeled sea-salt NO 3 fraction was lower than the D-NO 3. Mass concentration of fine and coarse NO 3 was in the same order of around 4 6 µg m 3 during the A1 and A2 periods. The formation of D-NO 3 is triggered by the uptake of HNO 3 gas during dust transport, which indicates that the coarse D-NO 3 is important for the prediction of fine NO 3 (for PM 2.5 too) and mass budget studies, including deposition to the ground surface. Particle volume concentration as a function of particle diameter and DR (DR-Dp-Volume plot) during the A1 and A2 periods is of interest. Figure 5 shows a snapshot during each dust period when the dust concentration peaked. The numbers in the figure show the 12-h averaged concentration from ACSA-12 and the FP/ IC analysis. The DR-Dp-Volume plot is divided into four sections according to Pan et al. (2016). A DR below 0.1 can be considered to be a spherical particle. Regions II, III, and IV can be considered to be anthropogenic aerosol particles, sea salt, and mineral dust, respectively. During the A2 period, region IV had the highest concentration due to dust onset. A small peak in the sea salt particle diameter was observed between 2 3 µm (Region III). This shows that the anthropogenic aerosol was also higher than during the A1 period (which had a good correspondence with the ACSA concentration). The POPC shown in Fig. 5 well supports ACSA-12 and FP/IC observations and GEOS-Chem model results. Fig. 5. Snapshot of particle volume concentration (color) as a function of particle diameter and depolarization ratio DR (DR-Dp-Volume plot) for periods (a) A1 and (b) A2. Numbers in the figure indicate the 12-h averaged observed aerosol concentration. 3.3 Concentration changes along the air mass trajectories Concentration changes during dust transport are important. Figures 6a and 7a show the HYSPLIT back trajectory. Back trajectories were calculated from 500-, 1000-, and 2000-m levels. Figures 6b and 7b show the simulated total dust, fine NO 3, total D-NO 3, HNO 3, and RH at the model first level (z = 130 m) along the HYSPLIT back-trajectory path (1000-m level path) for the A1 and A2 dust periods, respectively. During transport in period A1 (Fig. 6), the trajectory indicated that the air mass foot prints were similar at three levels. The lower level air mass passed over Fukuoka - East China Sea Shanghai - Jiangsu Province - the western edge of the North China Plain - Inner Mongolia (transport time from the Chinese coast to Fukuoka was about 30 h), and appears to have been influenced by the large levels of air pollution over the central-north China regions. Fine A-NO 3 concentrations peaked when the air mass left the Chinese mainland. After leaving the Chinese mainland toward the East China Sea, the NH 3 supply stopped and the thermal decomposition of NH 4 NO 3 provided additional HNO 3 gas that was available for uptake by dust, which was indicated by the increased concentration of D-NO 3. Over the ocean, the RH was higher and the HNO 3 concentration increased; therefore D-NO 3 formation continued (increasing ~4 µg/m 3 /day), even with a rapid reduction in dust concentration. When the air mass arrived in Fukuoka, the D-NO 3 concentration corresponded to 60% of the fine A-NO 3. This is important for mass budget studies of atmospheric nitrate. Transport in period A2 was complicated. Due to the vertical wind shear, the lower and upper transport paths were different. The lower air mass path was Fukuoka East China Sea Qingdao/ Shandong Province Anhui Province Janxi Province (the trans port time from the Chinese coast to Fukuoka was about 24 h), and the air mass was strongly influenced by the large area of

5 SOLA, 2017, Vol. 13, 47 52, doi: /sola Fig. 6. (a) Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) back trajectory calculated from Fukuoka at heights of 500, 1000, and 2000 m. Symbols are inserted at 6-h intervals, (b) simulated total dust, fine A-NO 3, total D-NO 3, HNO 3, and RH at the model first level (z = 130 m) along the HYSPLIT back-trajectory path (1000-m level path) for the A1 dust event. pollutant emissions in China. The upper air mass contained the pure dust. When the air mass trajectory passed over the daytime land area, dust aloft was entrained by convective mixing into the lower atmosphere (the decrease in RH indicates the location of dust entrainment), where it was mixed with anthropogenic pollutants. The first mixing took place over the Shandong Peninsula East China Sea, which was where D-NO 3 formation started. The second mixing and formation of D-NO 3 occurred over southern Korea. The fine A-NO 3 level over the ocean was lower than in the A1 period. The dust concentration rose and fell rapidly; therefore, the formation of D-NO 3 also followed the same pattern before it reached Kyushu. However, the total D-NO 3 level at Kyushu was higher than in period A1 (about 80% higher), because the RH and dust concentration were higher than during period A1. 4. Conclusions Heavy dust storms occurred between March 13 and 23, 2014 were analyzed by dust and anthropogenic aerosol measurement at 1-h intervals in Fukuoka and numerically simulated the Asianscale transport process. The conclusions of our study are summarized below. 1) Dust in the first period (A1) was transported without a cold front and was well mixed with anthropogenic pollutants. The numerical analysis explained this dust and pollution phenomena. 2) Dust in the second period (A2) was generated by a typical coldfront outbreak, and Chinese anthropogenic pollutants were transported within the warm sector of a low-pressure system ahead of the dust, while the dust was transported aloft behind a cold front. Air mass trajectories were different for the lower and upper layers. Dust was entrained into the lower layer and intermittently mixed with pollutants during transport to Korea and Japan. 3) In both cases, we observed an increase of coarse-mode nitrate with dust. The numerical model, including the heterogeneous dust acid uptake processes, explained this increase. We showed that the formation of dust-nitrate started over the Yellow Sea and East China Sea when the RH increased, and extra HNO 3 was supplied from the decomposition of NH 4 NO 3. We showed that more than 40% of nitrate exists in dust-nitrate when air mass arrived in Japan. Acknowledgements This work was supported by MEXT/JSPS KAKENHI Grants JP and JP16H The authors thank Dr. D. Fairlie of NASA for providing the dust acid uptake version of GEOS-Chem and Dr. S. S. Lee of KMA for supporting the PM observation data in Korea. Edited by: Y. Miyazaki Supplement Fig. 7. Same as Fig. 6, but for the A2 dust event. Supplement 1 describes the chemistry and heterogeneous reactive uptake calculation in GEOS CHEM.

6 52 Uno et al., Simultaneous Dust and Pollutant Transport over East Asia References Bey, I., J. Jacob, R. M. Yantosca, A. Logan, B. D. Field, A. M. Fiore, Q. Li, H. Y. Liu, J. Mickley, and M. G. Schultz, 2001: Global modeling of tropospheric chemistry with assimilated meteorology: Model description and evaluation. J. Geophys. Res., 106, Fairlie, T. D., D. J. Jacob, and R. J. Park, 2007: The impact of transpacific transport of mineral dust in the United States. Atmos. Environ., 41, Fairlie, T. D., D. J. Jacob, J. E. Dibb, B. Alexander, M. A. Avery, A. van Donkelaar, and L. Zhang, 2010: Impact of mineral dust on nitrate, sulfate, and ozone in transpacific Asian pollution plumes. Atmos. Chem. Phys., 10, Huang, R.-J., and co-authors, 2014: High secondary aerosol contribution to particulate pollution during haze events in China. Nature, 514, Kanatani, K. T., and co-authors, 2010: Desert dust exposure is associated with increased risk of asthma hospitalization in children. Amer. J. Respiratory Critical Care Medicine, 182, Kimoto, H., A. Ueda, K. Tsujimoto, Y. Mitani, Y. Toyasaki, and T. Kimoto, 2013: Development of continuous dichotomous aerosol chemical speciation analyzer. Clean Technol., 23, (in Japanese). Kobayashi, H., M. Hayashi, K. Shiraishi, Y. Nakura, T. Enomoto, K. Miura, H. Takahashi, Y. Igarashi, H. Naoe, and N. Kaneyasu, 2014: Development of a polarization optical particle counter capable of aerosol type classification. Atmos. Environ., 486, , doi: /j.atmosenv Kurokawa, J., T. Ohara, T. Morikawa, S. Hanayama, J.-M. Greet, T. Fukui, K. Kawashima, and H. Akimoto, 2013: Emissions of air pollutants and greenhouse gases over Asian regions during : Regional Emission inventory in ASia (REAS) version 2. Atmos. Chem. Phys., 13, Mori, I., M. Nishikawa, T. Tanimura, and H. Quan, 2003: Change in size distribution and chemical composition of kosa (Asian dust) aerosol during long-range transport. Atmos. Environ., 37, Pan, X. L., I. Uno, Y. Hara, K. Osada, S. Yamamoto, Z. Wang, N. Sugimoto, H. Kobayashi, and Z. F. Wang, 2016: Polarization properties of aerosol particles over western Japan: Classification, seasonal variation, and implications for air quality. Atmos. Chem. Phys., 16, Park, R. J., D. J. Jacob, B. D. Field, R. M. Yantosca, and M. Chin, 2000: Natural and transboundary pollution influences on sulphate-nitrate-ammonium aerosols in the United States: Implications for policy. J. Geophys. Res., 109, D15204, doi: /2003jd Shimizu, A., N. Sugimoto, and I. Matsui, 2010: Detailed description of a data processing system for lidar network in East Asia. Proceeding of 25th International Laser Radar Conference, St. Petersburg, Russia, Song, C. H., and G. R. Carmichael, 2001: A three-dimensional modeling investigation of the evolution processes of dust and sea-salt particles in east Asia. J. Geophys. Res., 106, Sugimoto, N., I. Matsui, A. Shimizu, T. Nishizawa, Y. Hara, C. Xie, I. Uno, K. Yumimoto, Z. Wang, and S.-C. Yoon, 2008: Lidar network observations of tropospheric aerosols. SPIE 7153, 2008, doi: / Uno, I., K. Osada, K. Yumimoto, S. Itahashi, S. Kuwahara, X. Pan, Y. Hara, and S. Yamamoto, 2016: Numerical analysis of trans-boundary transport of dust-nitrate during the longlasting yellow sand episode observed over the Northern Kyushu area in late May early June Atmos. Environ. Soc. Japan, 51, (in Japanese). Uno, I., N. Sugimoto, A. Shimizu, K. Yumimoto, Y. Hara, and Z. Wang, 2014: Record heavy PM2.5 air pollution over China in January 2013: Vertical and horizontal dimensions. SOLA, 10, , doi: /sola Yumimoto, K., I. Uno, X. Pan, T. Nishizawa, S. W. Kim, and N. Sugimoto, 2017: Inverse modeling of Asian dust emission with POPC observations: A TEMM dust sand storm 2014 case study. SOLA, 13, 31 35, doi: /sola Zhang, X. Y., S. L. Gong, Z. X. Shen, F. M. Mei, X. X. Xi, L. C. Liu, Z. J. Zhou, D. Wang, Y. Q. Wang, Y. Cheng, co-authors, 2003: Characterization of soil dust aerosol in China and its transport and distribution during 2001 ACE-Asia: 1. Network observations. J. Geophys. Res., 108. doi: / 2002JD Manuscript received 18 December 2016, accepted 20 February 2017 SOLA: jstage. jst. go. jp/browse/sola/