Transboundary Transport of Nitrogen Oxides from the Asian Continent to Fukue Island , Japan : Analyses of Long-Range Transport of Nitrogen Compounds

To investigate long-range transport and oxidation states of nitrogen compounds from the Asian continent, especially of total odd nitrogen species (NOy), we carried out continuous observations of NOy, total nitrate (T.NO3; the sum of gaseous nitric acid and particulate nitrate) and NOx (= NO + NO2) at Fukue Island, located in westernmost Japan. NOy and T.NO3 exhibited similar seasonal cycles, with maximum concentrations observed during winter-spring seasons, and minimum concentrations during summer. NOx had a seasonal cycle, with maximum and minimum concentrations in winter and summer, respectively; NOx concentrations also decreased markedly from winter to spring. High-concentration event analyses to assess transboundary pollution of NOy were performed. Transboundary pollution events for NOy were extracted, and classified as either “Case 1” or “Case 2”. Case 1 involved transport of NOy, T.NO3, and NOx from the Asian continent to Fukue Island, while Case 2 involved transport of NOy and T.NO3, but not NOx. Case 1 and Case 2 occurred predominantly in winter and spring, respectively. Air mass trajectories indicated that in Case 1, 52% of air masses passed through Korea, while only 5% passed through Korea in Case 2. These results indicate that some NOx emitted from the Asian continent is transported to Fukue Island as NOx, without it undergoing oxidation, reflecting low photochemical activity and/or short transport times.


INTRODUCTION
In East Asia, emission of nitrogen oxides (NO x = NO + NO 2 ) has increased, related to the marked recent economic growth.A recent emission inventory for Asia showed that NO x emissions in Asia and China have increased by 54% and 89%, respectively, over the period from 2000 to 2008 (Kurokawa et al., 2013).NO x emitted in East Asia is transported long range, undergoing chemical transformations that directly and indirectly affect environmental quality in other regions.Typically, concentrations of the total odd nitrogen species (NO y ) are assessed, instead of NO x , to evaluate long-range transport of nitrogen compounds.NO y consists of NO x , NO 3 , N 2 O 5 , peroxyacyl nitrates, organic nitrates, nitrous acid, gaseous nitric acid (HNO 3 ), particulate nitrate (NO 3 -(p)), and other minor compounds (e.g., ClNO 2 ).NO y is removed from the atmosphere through deposition to ground or sea surfaces, as HNO 3 and NO 3 -(p), where it plays an important role in the nitrogen cycle in Earth systems * Corresponding author.
Study on the long-range transport of NO y from the Asian continent has usually focused on HNO 3 and NO 3 -(p).Takiguchi et al. (2008) measured NO y , HNO 3 , and NO 3 -(p) at Cape Hedo, Okinawa, Japan.They showed that the ratio of T.NO 3 (= HNO 3 + NO 3 -(p)) to NO y rose, with increasing transport time from China to Cape Hedo.In addition, NO 3 -(p) was mainly found in the coarse aerosol mode at Cape Hedo.Simultaneous observations of NO y and T.NO 3 at Cape Hedo and Fukue Island, Nagasaki, Japan, indicated that the ratio of T.NO 3 to NO y increased during transport from Fukue Island to Cape Hedo, especially in fresh air masses (i.e., initially low T.NO 3 /NO y ratios) (Yuba et al., 2014).Fujiwara et al. (2014) performed aerial observations of NO y , HNO 3 , and NO 3 -(p) over the East China Sea in autumn and winter.They showed that NO 3 -(p)/T.NO 3 ratios were less than 0.5 in most cases, suggesting that the main formation process of NO 3 -(p) was through heterogeneous reaction of HNO 3 on sea salt surfaces, except during Asian dust events.
The main formation process of T.NO 3 is through the reaction of NO 2 with an OH radical (Sadanaga et al., 2006): where M represents a third body in the reaction (mainly N 2 and O 2 in the atmosphere).Reaction (1) occurs during the daytime, while T.NO 3 is produced via the following reactions during the night (Matsumoto et al., 2006;Bertram and Thornton, 2009): In a clean atmosphere, N 2 O 5 concentrations are low, and Reaction (1) provides the main source of T.NO 3 (Yuba et al., 2014).In this case, NO x oxidation and T.NO 3 formation occur mainly via photochemical reaction in the presence of solar radiation during long-range transport.In fact, the aerial observations over the East China Sea showed that T.NO 3 /NO y ratios in autumn were higher than in winter (Fujiwara et al., 2014).This reflects the fact that solar radiation in winter is weaker, yielding lower photochemical activity, than in autumn.Low photochemical activity results in slower oxidation reaction rates of NO x .The aerial observations over the East China Sea suggested that NO x originating from the Asian continent remained unoxidized over the East China Sea in winter (Fujiwara et al., 2014).Continuous measurements of NO 2 at altitudes of 0-1 km using Multi-Axis Differential Optical Absorption Spectroscopy at Cape Hedo showed that NO 2 concentrations increased, when air masses were transported from China, within a period of about 24 h (Takashima et al., 2011).These results indicate that NO x emitted from the Asian continent can be transported to western Japan as NO x , without undergoing oxidation.
Determining the oxidation state of NO y is clearly one of the important factors involved in understanding the longrange transport of nitrogen compounds from the Asian continent to Japan.In fact, physical and chemical properties of T.NO 3 are different from those of NO x and their effects on the Earth's environment are different.In addition, the oxidation rate of NO x is related to the lifetime of NO y in the atmosphere, and determines the extent of the transboundary air pollution of NO y derived from NO x emitted on the Asian continent.
In this study, we carried out continuous observations of NO y , T.NO 3 , and NO x at a site on Fukue Island, Nagasaki, located at one of the points closest to the Asian continent in Japan.In order to investigate and clarify oxidation states of NO y transported from the Asian continent, especially, the transboundary transport of unoxidized NO x , we investigated variations in concentrations of these species during transboundary transport from the Asian continent.In particular, transboundary pollution events, involving NO y were extracted, and investigated in more detail.

OBSERVATIONS
Observations were carried out at the National Institute for Environmental Studies' Fukue Atmosphere Observation site (Fukue site; 32.8°N, 128.7°E) on Fukue Island, Nagasaki, Japan (Fig. 1).Details of the Fukue site are described elsewhere (e.g., Takami et al., 2013;Irei et al., 2016).Briefly, Fukue Island is situated at the western end of Japan, within the direct pollution outflow from the Asian continent.The population of Fukue Island is about 40,000 and there are no major industrial areas on the island.The observation site is located on the northwestern side of the island, on the upstream side of the outflow from the Asian continent.Therefore, the observation site is ideal for observing outflow from the Asian continent.Observations were carried out for a 2-year period, covering 2012 and 2013.
NO y , T.NO 3 , and NO x concentrations were continuously observed in this study.NO y and T.NO 3 concentrations were measured using a scrubber-difference NO-O 3 chemiluminescence (SD-CL) method.Details of the SD-CL method are described elsewhere (Sadanaga et al., 2008a, b;Yuba et al., 2010).Briefly, the SD-CL method has two lines, an "NO y line" and an "NO y -T.NO 3 line".On the NO y line, ambient air is introduced to a molybdenum reducing catalyst heated at 598 K (Mo catalyst), reducing NO y to NO; the NO concentration was then measured by an NO-O 3 chemiluminescent detector (Model 42i-TL; Thermo Fisher Scientific Inc., Waltham, MA, USA).NO y concentrations were measured using the signal from the NO y line.On the NO y -T.NO 3 line, ambient air was passed through a Teflon filter and an annular denuder coated with NaCl to remove NO 3 -(p) and HNO 3 , respectively, before being introduced to another Mo catalyst to reduce nitrogen species, which were then measured by the NO-O 3 chemiluminescent detector.NO y -T.NO 3 concentrations were measured using the signal from the NO y -T.NO 3 line.T.NO 3 concentrations are obtained by subtracting the simultaneous NO y -T.NO 3 concentration from the NO y concentration: (5) NO x concentrations were measured using a light-emitting diode photolytic converter, in combination with the NO-O 3 chemiluminescence method (Sadanaga et al., 2010).The detection limits of NO x and NO y were 60 pptv, with a 1-min integration time (2σ).The detection limit of T.NO 3 depends on the NO y concentration, and was estimated to be 71 pptv, with a 10-min integration time (2σ), under an NO y concentration of 5 ppbv (Sadanaga et al., 2008a).
Backward trajectory analyses were performed using the HYSPLIT 4 model developed by the American National Oceanic and Atmospheric Administration (NOAA) (Stein et al., 2015;Rolph, 2016).Initial altitude and calculation time were set to 500 m and 120 h, respectively.Origins of air masses reaching the observational site were classified into seven groups, based on the last coastline they passed.Air masses originated from North China (NC), North China and Korea (NCK), Korea (KR), South China (SC), Japan (JP), Ocean (O), and Russia (RU) (Fig. 1).NC, (NCK + KR), SC, JP, O and RU were divided according to the dashed lines drawn from the site in Fig. 1, and then NCK and KR were distinguished by the trajectories.NCK indicates air masses passed through metropolitan areas of North China (e.g., Beijing and Tianjin) and Korea, and contained air pollutants from both North China and Korea (Fig. 1).KR represents air masses that passed through Korea, but not through the metropolitan areas of North China.It should be noted that air masses originating from KR can pass through Northeast China (e.g., Harbin and Changchun), and may contain air pollutants from both Korea and Northeast China, but not from metropolitan areas of North China, such as Beijing and Tianjin (Fig. 1).Air masses that meandered, or did not belong to any of these categories were excluded from our analysis.Trajectory data were collated every 6 h, at 3:00, 9:00, 15:00 and 21:00 Japan Standard Time (JST).NO x , NO y , and T.NO 3 concentrations were averaged into 6-h bins to relate them to air masses identified using backward trajectory analysis.Thus, concentration data for time intervals 0:00-6:00, 6:00-12:00, 12:00-18:00 and 18:00-24:00, were equated with trajectories for 3:00, 9:00, 15:00, and 21:00, respectively.The numbers of each trajectory sector in each month during the observation period are shown in Supplementary Table S1.Air masses from ocean (i.e., O) and China (i.e., NCK and NC) reached Fukue Island frequently in summer and other seasons, respectively.The air mass frequencies from KR, SC, and RU to Fukue Island were low.

Seasonal Variations of Total Odd Nitrogen Species and its Constituents
Fig. 2 shows seasonal variations of monthly averaged NO y , T.NO 3 , and NO x concentrations.Daily averaged concentration variations of NO y , T.NO 3 , and NO x are shown in Supplementary Fig. S1.The missing data of NO y and T.NO 3 in October 2012 were due to a trouble of the NO-O 3 chemiluminescent detector used in the NO y and T.NO 3 Fig. 2. Variations in monthly averaged NO y , T.NO 3 , and NO x concentrations over a 2-year period, beginning January 2012 to December 2013.measurement system.Average concentrations of NO y , T.NO 3 , and NO x during the observation period were 3.11, 1.19, and 1.20 ppbv, respectively.NO y and T.NO 3 exhibited similar seasonal cycles, with maximum concentrations in winter-spring seasons, and minimum concentrations in summer.Seasonal cycles of NO y and T.NO 3 in 2013 had clearer peak and dip than those in 2012.This difference would be because the number of the transport events in 2013 was more than that in 2012 (see Fig. 5).The seasonal variation of NO x concentrations was slightly different from NO y and T.NO 3 concentrations.NO x had a seasonal cycle, with maximum and minimum concentrations in winter and summer, respectively.NO x concentrations also decreased markedly from winter to spring.
The variations in NO y , T.NO 3 , and NO x concentrations associated with the seven different air mass origins, based on backward trajectory analyses, are shown in Supplementary Fig. S2.Typically, NO y and T.NO 3 concentrations from NC and NCK were high, while those from JP and O were low.In contrast, the concentrations from JP, NC, and NCK air mass origins were all high for NO x .This is because air masses from JP can pass through source regions of NO x , such as urban areas in Kyushu Island, Japan, relatively close to Fukue Island.Fig. 3 shows variations in NO y , T.NO 3 , and NO x concentrations in NC and NCK air masses, for which NO y , T.NO 3 , and NO x concentrations were high.In winter, NO x concentrations in air masses that passed through Korea (i.e., NCK) were significantly higher than in those from NC.This pattern was not observed for NO y and T.NO 3 concentrations, or for NO x concentrations in other seasons.This suggests that some of the NO x emitted from the Asian continent, especially from Korea, reached Fukue Island, without undergoing oxidation in winter.In other seasons, almost all the NO x from the Asian continent reacted to form descendant photochemical products, such as T.NO 3 , before arriving at Fukue Island.

Transport Event Analyses of Total Odd Nitrogen Species from the Asian Continent
To investigate transboundary transport of NO x , NO y , and T.NO 3 from the Asian continent to Fukue Island, event analyses were performed.Transport events were extracted using variations in concentrations of NO y and T.NO 3 .These events occurred when concentration peaks for both NO y and T.NO 3 continued for more than 12 h; and when the origins of air masses arriving at Fukue Island were from the Asian continent (i.e., having NC, NCK, KR or SC air mass origins).We identified the increase in concentrations of NO y and T.NO 3 during such events, as the result of transboundary transport of nitrogen compounds from the Asian continent.
We identified two kinds of the transboundary transport events, defined here as "Case 1" and "Case 2".Fig. 4 shows examples of Case 1 and Case 2 events.In Case 1, NO x , NO y , and T.NO 3 concentrations all increased (Fig. 4(a)), indicating that NO y , T.NO 3 , and NO x were all transported from the Asian continent to Fukue Island.In Case 2, NO y and T.NO 3 concentrations increased, but the NO x concentration did not (Fig. 4(b)).This implies NO y and T.NO 3 were transported from the Asian continent, but NO x was thoroughly consumed before arriving at the site.
Cases 1 and 2 were extracted based on variations in concentrations of NO y , T.NO 3 , and NO x for 2012 and 2013.Integration times for these transboundary transport events were calculated using 1-h averaged NO y concentrations.Integration time is shown in Fig. 4 and was defined as a peak width at baseline.Fig. 5 shows monthly variations of integration times for Cases 1 and 2 for 2012 and 2013.In  winter (December, January, and February), Case 1 accounted for 74% of events, while Case 2 accounted for 86% of events in spring (March, April, and May).These results indicate that NO y and T.NO 3 were transported from the Asian continent to Fukue Island in spring, whereas NO x , NO y , and T.NO 3 were transported simultaneously in winter.Differences in air mass origins between Cases 1 and 2 also were investigated.Fig. 6 shows monthly variations in integration times reflecting air mass origins for Cases 1 and 2 for 2012 and 2013.Air mass trajectories indicate that in Case 1, 52% of air masses passed through Korea (i.e., having NCK or K air mass origins), while only 5% passed through Korea in Case 2 (i.e., they had NC or SC air mass origins instead).Korea is geographically closer to Fukue Island than China, and transport times from the Asian continent to Fukue Island could affect transboundary transport of NO x .Fig. 7 shows monthly variations in transport times from the Asian continent to Fukue Island for Cases 1 and 2. The transport time was defined as the time taken for air masses to travel from a coastline on the Asian continent to the Fukue site; thus, it reflects the time spent over the sea,  calculated using the backward trajectory analysis.The transport times for Case 1 were clearly shorter than for Case 2 events, yielding average transport times for Case 1 and 2 of 17.8 h and 36.6 h, respectively.

General Discussion
As described in the previous sections, some of the NO x emitted from the Asian continent reached Fukue Island in the form of NO x in winter, without undergoing oxidation; this did not occur in other seasons.In addition, it was mainly air masses passing through Korea on route to Fukue Island that transported NO x emitted from the Asian continent, without it undergoing oxidation.These results suggest that low photochemical activity and short transport times from the Asian continent to the Fukue site are crucial factors determining NO x transport.The oxidation of NO x to T.NO 3 proceeds via photochemical reactions under solar radiation.Thus, the lifetime of NO x is short, in the case of strong solar radiation and high photochemical activity.In the present study, the solar radiation and photochemical activity in winter are lower than in spring; thus, the lifetime of NO x is longer and NO x emitted from the Asian continent is able to arrive at Fukue Island in the form of NO x .Air mass origins and transport times also pay a role, since Korea is geographically closer to Fukue Island than China, and transport times of air masses from NCK or K are shorter than from NC or SC.The shorter transport time implies a shorter reaction time for NO x oxidation during transport from the Asian continent to the Fukue site; hence, NO x emitted from the Asian continent can arrive at Fukue site, before being oxidized.
The main oxidation of NO x to T.NO 3 during transport is via the reaction of NO 2 with an OH radical (Reaction (1)).The rate constant for this reaction has been reported as 1.13 × 10 -11 cm 3 molecule -1 s -1 under normal pressure and temperature of 1013 hPa and 288 K (Burkholder et al., 2015).The diurnally averaged concentration of OH radicals in the atmosphere is 0.91-1.37× 10 6 radicals cm -3 in remote areas (Monks et al., 1998;Smith et al., 2006;Kanaya et al., 2007).The lifetime of NO x is calculated to be 18-27 h.Given the average transport times for Case 1 and 2 of 17.8 h and 36.6 h obtained in the previous section, NO x emitted from the Asian continent only has enough time to be oxidized to T.NO 3 in Case 2. In Case 1, the average transport time is shorter than the NO x lifetime, and some of the NO x emitted from the Asian continent will reach Fukue Island as NO x .

CONCLUSIONS
Continuous observations of NO y , T.NO 3 , and NO x were carried out at Fukue site, located in the westernmost Japan, in order to investigate oxidation states of NO y transported from the Asian continent.Seasonal variations and transport events of NO y , T.NO 3 , and NO x were analyzed, and we led to the following conclusions.Some of the NO x emitted from the Asian continent was transported to Fukue Island as NO x , in winter, when there was low photochemical activity.Under low photochemical activity, the lifetime of NO x is longer, and NO x emitted from the Asian continent is able to reach Fukue Island as NO x .NO x was likely to be transported from the Asian continent to Fukue Island as NO x when there was a short transport time.A short transport time results in a short reaction time for NO x oxidation to occur during transport from the Asian continent to the Fukue site, allowing NO x emitted from the Asian continent to arrive at the Fukue site before being oxidized.Previous research (Takashima et al., 2011;Fujiwara et al., 2014) also suggests the long-range transport of unoxidized NO x from the Asian continent, and transboundary pollutions of nitrogen compounds from the Asian continent, including various oxidation states of NO y , are important contributions to nitrogen pollution in East Asia.

Fig. 1 .
Fig. 1.Map of East Asia and location of the Fukue site showing classification of air masses based on origins: North China (NC), North China and Korea (NCK), Korea (KR), South China (SC), Japan (JP), Ocean (O), and Russia (RU).

Fig. 3 .
Fig. 3. Variations in monthly averaged NO y , T.NO 3 , and NO x concentrations in air masses from North China (NC) and North China and Korea (NCK) over a 2-year period, beginning January 2012 to December 2013.

Fig. 4 .
Fig. 4. Examples of the transport events for (a) Case 1 (December 2012) and (b) Case 2 (May 2013).Solid, gray, and dashed lines indicate NO y , T.NO 3 , and NO x concentrations, respectively.Each plot shows 1-h averages.

Fig. 5 .
Fig. 5. Monthly variations in the integration times for Cases 1 and 2 over a 2-year period, beginning January 2012 to December 2013.

Fig. 6 .
Fig. 6.Monthly variations in the integration times, classified according to air mass origin for Cases 1 and 2 over a 2year period, beginning January 2012 to December 2013.

Fig. 7 .
Fig. 7. Monthly variations in transport times from the Asian continent to the Fukue site for Cases 1 and 2 over a 2-year period, beginning January 2012 to December 2013.