HTML
--> --> -->Since wintertime tropical intraseasonal oscillation (ISO), known as Madden-Julian Oscillation (MJO), was described in the early 1970s (Madden and Julian, 1971, 1972), studies (Madden, 1986; Wang and Rui, 1990; Zhang and Dong, 2004; Kikuchi et al., 2012) have proven that ISO exists not only in winter but also in summer. In summer, ISO has prominently different characteristics when compared to wintertime MJO. Wintertime MJO is represented by tropical convection, which is generated over the equatorial Indian Ocean and propagates eastward at speeds of ～5 m s-1 to the western Pacific Ocean in 30-60 days (Madden and Julian, 1994; Zhang, 2005; Lau and Waliser, 2012). After this tropical convection propagates to the date line, it gradually weakens and finally disappears (Lau and Chan, 1985; Knutson and Weickmann, 1987; Wang and Rui, 1990). Compared to wintertime MJO, the tropical deep convection related to boreal summer intraseasonal oscillation (BSISO) shows northward propagation (Lee et al., 2013) in addition to eastward propagation. This northward propagation is considered closely related to the summer monsoon (Yasunari, 1979; Li and Wu, 2000; Annamalai and Slingo, 2001; Chan et al., 2002; Han et al., 2006; Wen and Zhang, 2008). In terms of strength, BSISO is generally weaker than wintertime MJO (Madden and Julian, 1994; Wheeler and Hendon, 2004; Zhang, 2005).
As the 30-60-day BSISO propagates northeastward across the Indian summer monsoon (ISM) region, signals of intraseasonal variation are observed along the propagation pathway from the equatorial Indian Ocean to the Northwest Pacific (Yasunari, 1979; Lau and Chan, 1986; Lawrence and Webster, 2002). Strong ISO signals in the ISM region show up as copious/scant rainfall occurring in India, called active/break spells of the ISM (Raghavan, 1973; Waliser, 2006; Rajeevan et al., 2010). BSISO has been proven to be closely related to summer monsoon onset (Kang et al., 1999; Lee et al., 2013) and the active/break periods of summer monsoon (Annamalai and Sperber, 2005; Wang et al., 2005).
Studies have revealed that wintertime MJO affects tropospheric weather (Lau and Chan, 1986; Matthews et al., 2004; Donald et al., 2006; Pohl et al., 2007, Pohl et al., 2010) and stratospheric polar vortex variation, possibly even causing sudden stratospheric warming events (Garfinkel et al., 2012; Liu et al., 2014). Additionally, since the MJO signal in total column ozone (TCO) was first observed in Nimbus-7 Total Ozone Mapping Spectrometer data (Sabutis et al., 1987; Gao and Stanford, 1990), the wintertime MJO has, through multiple other observations, also been proven to have an influence on atmospheric ozone. Due to the MJO being active from the equatorial Indian Ocean to the western Pacific, the intraseasonal variability of TCO is mostly obvious in the subtropics over the Eastern Hemisphere and the Pacific (Tian et al., 2007; Lau et al., 2012). The eastward-propagating MJO in the tropics causes a series of ozone minima events over the Tibetan Plateau in winter (Liu et al., 2009, 2010a). Recently, the most significant wintertime intraseasonal ozone variability related to the MJO was found in the subtropical upper troposphere-lower stratosphere (UTLS), where anticyclonic/cyclonic circulation anomalies are generated by equatorial anomalous convective forcing. These circulation anomalies in the subtropical UTLS result in variation of the tropopause height, which initiates the ozone-column variation in this region (Tian et al., 2007; Li et al., 2012; Liu et al., 2015a; Zhang et al., 2015).
BSISO propagation is of importance in modulating circulation, thus affecting weather systems and even tropical cyclones (Bessafi and Wheeler, 2006; Kikuchi and Wang, 2010). In this paper, we try to reveal the influence of BSISO on tropospheric ozone in the ISM region, especially during extreme active and break periods of the ISM. We also find that precipitation and horizontal transport play important roles in the BSISO of lower-tropospheric ozone. Section 2 describes the data and methods used. In section 3, we present the general features of ozone anomalies caused by BSISO, and their correlations with precipitation and horizontal transport. In section 4, we further compare the intraseasonal variability of lower-tropospheric ozone during active and break spells of the ISM, before drawing conclusions in section 5.
2.1. Data
The daily mean ozone, zonal wind and meridional wind are from the ERA-Interim dataset (e.g., Dee et al., 2011). The climatology of the lower-tropospheric ozone column is from Ozone Monitoring Instrument (OMI) data. OMI ozone profiles in the troposphere have been validated (Liu et al., 2010b) and used in studying lower-tropospheric ozone (Lu et al., 2018). The precipitation related to BSISO is based on TRMM observations (Kummerow et al., 2000). The daily mean outgoing longwave radiation (OLR) data are from the AVHRR instrument onboard NOAA's polar-orbiting spacecraft (Liebmann and Smith, 1996). The time period of these datasets considered in this study is summer (June-August) 2000-2012.It has been proven that BSISO has quasi-oscillating periods of 30-60 days (Lee et al., 2013). As a result, a 30-60-day bandpass filter is applied to the daily anomalies (to remove the climatology from the daily mean) of variables like OLR, horizontal wind and ozone to derive their BSISO characteristics (i.e., BSISO-related OLR, horizontal wind and ozone anomalies). However, as bandpass filtering will reduce the effective sample size, the regular Student's t-test is no longer capable of testing the significance of anomalies related to BSISO (e.g., Tian et al., 2011; Liu et al., 2014). Instead, we use a two-tailed Student's t-test with reduced degrees of freedom (Tian et al., 2011; Liu et al., 2014; Zhang et al., 2015) to examine the BSISO-related composite results in this work.
We use ozonesonde profiles at three stations (Fig. 1) from the World Ozone and Ultraviolet Radiation Data Centre (WOUDC). Hanoi (AAR) station (21.01°N, 105.8°E) is located in northern Vietnam, nearly 90 km away from the coast. Ozonesonde observations have been made every two or three weeks since 2004, with a gap in 2011 and 2012. Sepang Airport (SEP) station (2.73°N, 101.7°E) in Singapore has a long history of ozonesonde observations, every week or every two weeks, since 1998. Naha (NAH) station (26.21°N, 127.69°E) is located in southernmost Japan, observing ozone profiles since 1989.
Figure1. (a) Climatology of the lower-tropospheric ozone column (700 hPa to the surface) in June-July-August, based on OMI observations. The red dots are the three ozonesonde locations. The solid blue line is the northeastward transect, and the blue dashed line is the northward transect. (b) Variance of the BSISO-related lower-tropospheric ozone column (700 hPa to the surface) in June-July-August, based on ERA-Interim data.2
2.2. BSISO index
BSISO is more complicated than wintertime MJO, as it involves the northward propagation of deep convection extending further from the equator to the extratropical Northern Hemisphere. The widely-used real-time multivariate MJO index (Wheeler and Hendon, 2004) is unable to show this northward propagation. Therefore, we use another BSISO index, suggested by (Lee et al., 2013), which is based on multivariate empirical orthogonal function (MV-EOF) analysis of OLR and zonal wind at 850 hPa in the region of (10°S-40°N, 40°-160°E). The BSISO index is defined by the first two principal components (PCs) of the MV-EOF analysis, representing the northward and eastward propagation of the 30-60-day BSISO. The lifecycle of BSISO is divided into eight phases according to this BSISO index, indicating the location of deep convection along its northeast propagation pathway. We select BSISO events with amplitudes greater than 1.0 [(PC12 + PC22)1/2>1.0] in this study.2
2.3. Active/break events of the ISM
There are intraseasonal variabilities of convection and rainfall over the ISM region. ISM rainfall fluctuates between being copious and scant, referred to as active and break spells of the ISM, respectively. These two spell types are usually defined during the peak monsoon months (July and August) by the average daily rainfall over the core ISM area (18.0°-28.0°N, 65.0°-88.0°E). According to the definition, an active (break) spell of the ISM is a period of three consecutive days or more during which the rainfall anomaly over the core ISM region is more (less) than +1.0 (-1.0) the standard deviation (Rajeevan et al., 2010).3.1. Tropospheric ozone in the ISM region
Figure 1a shows the climatology of the lower-tropospheric ozone column (700 hPa to the surface) in summer (June-July-August) based on OMI data. The ozone in the midlatitudes is higher than in the tropics, with the most significant gradients at 25°-30°N. In the tropics, the ozone over the western Pacific is lower than over the Indian Ocean.After applying a 30-60-day bandpass filter to the lower-tropospheric ozone column, we obtain the variance of BSISO-related ozone in the lower troposphere (Fig. 1b). There are two areas in which the most significant variance is observed: one is the Maritime Continent, especially Malaysia and the Celebes Sea; and the other is the western North Pacific. The lower-tropospheric ozone in these two areas has the most prominent variability on the intraseasonal time scale. The power spectra of the lower-tropospheric ozone (Fig. 2) in the Maritime Continent (5°S-10°N, 100°-140°E) and western North Pacific (15°-30°N, 120°-140°E) confirm that the ozone variance is concentrated in intraseasonal periods of 30-60 days, especially in the Maritime Continent (Fig. 2a).
Figure2. Power spectra of BSISO-related anomalies of the lower-tropospheric ozone column (700 hPa to the surface) in the (a) Maritime Continent and (b) western North Pacific. The red curve is the red-noise spectrum. The lower and upper blue dashed curves are 5% and 95% red-noise significance levels respectively.2
3.2. Horizontal distribution and propagation of BSISO-related ozone
The lifecycle of BSISO is 30-60 days, as also reported by (Lee et al., 2013). They divided the BSISO's lifecycle into eight phases according to the location of deep convection. The composite 30-60-day bandpass-filtered (BSISO-related) OLR anomalies represent the location of deep convection in each BSISO phase (Figs. 3a-h). The enhanced/suppressed convection (negative/positive OLR anomalies), accompanied by a negative/positive lower-tropospheric ozone column (700 hPa to the surface) anomalies, show a northwest-southeast pattern, which propagates northeastward, during phases 1-8. The BSISO-related deep convection (negative OLR anomalies) generate over the equatorial eastern Indian Ocean in BSISO phase 1 and enhance in phase 2. Simultaneously, negative BSISO-related anomalies of ozone are observed over the eastern Indian Ocean and Maritime Continent. During phases 3-4, the BSISO-related convection weakens slightly and gradually propagates northeastward to India, the Bay of Bengal, and the Maritime Continent. The negative BSISO-related ozone anomalies over the eastern Indian Ocean and Maritime Continent are also reduced in this period. After phase 5, the negative BSISO-related OLR anomalies and the negative ozone anomalies gradually propagate to the western North Pacific, until the start of the next BSISO lifecycle. Similarly, suppressed convection (positive BSISO-related OLR anomalies) develops over the equatorial Indian Ocean in phase 4, and enhances during phases 5-7 with a northeastward propagation. The positive anomalies of BSISO-related ozone are observed over the eastern Indian Ocean and Maritime Continent during this period. As the suppressed convection propagates to the western North Pacific after phase 8, positive ozone anomalies also gradually propagate to the Northwest Pacific until the next phase 5, when positive ozone anomalies develop over the eastern Indian Ocean and Maritime Continent.
Figure3. Composites of BSISO-related (a-h) OLR anomalies (units: W m-2) and (i-p) lower-tropospheric column (700 hPa to the surface) ozone anomalies (units: DU). Positive and negative anomalies are indicated by red solid and blue dashed lines, respectively. The shaded areas are statistically significant at the 95% confidence level, based on a two-tailed Student's t-test with reduced degrees of freedom. The top-right number in each panel indicates the number of days used for each composite.However, there are slight inconsistencies between the locations of BSISO-related OLR anomalies and ozone anomalies. For example, when negative/positive OLR anomalies are in the eastern Indian Ocean in phases 1-2/5-6, the most significant negative/positive ozone anomalies are not, instead being situated in the Maritime Continent. These inconsistences indicate that there are more complicated factors influencing the BSISO of lower-tropospheric ozone. Thus, we investigate the factors that might influence the production and transport of lower-tropospheric ozone. Limited by the availability of data, we only find two factors that are closely related to the BSISO of lower-tropospheric ozone: precipitation anomalies and horizontal circulation anomalies. Precipitation will clean ozone precursors like volatile organic compounds in the troposphere and reduce temperature, resulting in a reduction of generated ozone. Horizontal circulation anomalies can transport ozone across the horizontal gradients of ozone, especially in the midlatitudes, where the most significant ozone gradients are observed (Fig. 1a). Furthermore, the composite BSISO-related precipitation anomalies, stream function anomalies, and horizontal wind anomalies (Fig. 4) show the important roles played by the precipitation and horizontal transport caused by BSISO in modulating the BSISO of lower-tropospheric ozone.
Figure4. As in Fig. 3 but for (a-h) TRMM precipitation anomalies (units: mm) and (i-p) stream function anomalies (contours; units: 106 m2 s-1) and horizontal wind anomalies (vectors; units: m s-1) at 850 hPa. The red vectors are statistically significant at the 95% confidence level, based on a two-tailed Student's t-test with reduced degrees of freedom.Compared to OLR, BSISO-related positive/negative precipitation anomalies are more consistent with negative/ positive ozone anomalies. In phases 1-3, positive precipitation anomalies extend form the eastern Indian Ocean to the equatorial Western Pacific, resulting in negative ozone anomalies over the Maritime Continent (Figs. 3i and k). At the same time, there is a cyclonic anomaly over the Maritime Continent and an anticyclonic anomaly over the north of the Maritime Continent (Figs. 4i-k). Strong easterly flow between them brings air containing low ozone (Fig. 1a) from the Pacific to the Maritime Continent, reducing the level of tropospheric ozone in the latter. Positive precipitation and anticyclonic anomalies gradually propagate northeastward to the western Pacific after phase 4. In the following phases 5-7, negative precipitation anomalies occur in the Maritime Continent (Figs. 4e-g), causing positive ozone anomalies (Figs. 3e-g). In this period, there is a cyclonic anomaly (Figs. 4m-o) north of the Maritime Continent, which is more powerful than the anticyclonic anomaly over the Maritime Continent. The northwesterly winds at the edge of the cyclone bring air containing more ozone (Fig. 1a) from mainland Southeast Asia to the Maritime Continent, increasing the tropospheric ozone in this area until phase 8, when the cyclonic anomaly propagates northeastward to the western North Pacific.
It is worth noting that when positive precipitation anomalies propagate to the western North Pacific in phase 6 (Fig. 4f), the negative ozone anomalies become too weak to be observed (Fig. 3n). However, in the following phases 7 and 8, negative ozone anomalies once again occur in the western North Pacific (Figs. 3o and p), which seems inconsistent with disappearing positive precipitation anomalies (Figs. 4g and h). This is because of the strengthened cyclonic anomalies over the western North Pacific (Figs. 4o and p). The southerly winds at the eastern edge of the cyclone bring air containing less ozone in the tropics to the north, reducing the tropospheric ozone in the western North Pacific. Similarly, the positive ozone anomalies in the western North Pacific in phases 3 and 4 (Figs. 3k and l) are caused by northerly winds at the edge of anticyclonic anomalies (Figs. 4k and l).
In order to elucidate the relative importance of precipitation and horizontal transport, we calculate the correlation between the BSISO-related ozone anomalies and the precipitation anomalies (Fig. 5a) as well as the stream function anomalies at 850 hPa (Fig. 5b). BSISO-related ozone is better correlated with precipitation anomalies in the eastern Indian Ocean and Maritime Continent; however, in the western North Pacific, ozone is better correlated with stream function anomalies. Based on the above analysis, one can conclude that precipitation plays a dominant role in modulating the BSISO of lower-tropospheric ozone in the tropics (eastern Indian Ocean and Maritime Continent). However, horizontal transport is relatively more important for the BSISO of lower-tropospheric ozone in the extratropics (western North Pacific).
Figure5. Correlation coefficients of BSISO-related anomalies of lower-tropospheric column (700 hPa to the surface) ozone with (a) precipitation and (b) 850-hPa stream function (only correlation coefficients greater than 0.5 are plotted).2
3.3. Vertical structure of BSISO-related ozone
To investigate the vertical structure of the BSISO-related ozone anomaly, we choose two transect locations: northward, almost across AAR station; and northeastward, across SEP station and close to NAH station (Fig. 1). Vertical cross sections of BSISO-related ozone anomalies and stream function anomalies along these two transect locations are compared in Fig. 6. In both cross sections, BSISO-related ozone anomalies are most remarkable in the tropics, where they extend from the surface to almost 300 hPa. Besides, the ozone anomalies in the tropics have a northward tilt as the height increases. When the ozone anomalies propagate to extratropical regions, they quickly weaken with a limited vertical range from the surface to about 700 hPa. As the BSISO-related anomalies continue to propagate northeastward, ozone anomalies are observed over the western North Pacific along the northeastward transect and South Asia along the northward transect.Along the northeastward transect, the nearest locations from SEP and NAH stations are marked as vertical red lines (Fig. 6). SEP station experiences the most significant negative/positive ozone anomalies in phases 1-2/5-6; however, in the subtropics, NAH station shows the most significant negative/positive ozone anomalies in phases 8-1/4-5. Along the northward transect, locations nearest to SEP and AAR stations are marked as vertical red lines (Fig. 6). AAR station experiences the most significant negative/positive ozone anomalies in phases 2/6. Therefore, we test ozone profiles in these BSISO phases based on ozonesonde data from SEP, NAH and AAR stations (Fig. 7). At SEP station, the ozonesonde data show larger ozone partial pressure in phases 5-6 than in phases 1-2 at almost every height from the surface to 500 hPa. This is consistent with the most significant positive ozone anomalies in phases 5-6 (Figs. 6e and f) and negative ones in phases 1-2 (Figs. 6a and b), which extend from the surface to 500 hPa in the tropics. When BSISO propagates northeastward to the western North Pacific, the most significant positive/negative ozone anomalies are in phases 4-5/8-1 (Figs. 6d, e, h and a). This is proved by the ozonesonde data from NAH station, which suggest that the lower-tropospheric ozone partial pressure is larger in phases 4-5 than in phases 8-1. At AAR station, the ozone partial pressure in phase 2 is larger than in phase 5 in the lower troposphere, which is also consistent with the positive ozone anomaly in phase 2 (Fig. 6j) and negative one in phase 6 (Fig. 6n).
Figure6. Vertical cross sections of BSISO-related ozone anomalies (units: DU km-1, colored contours) and stream function anomalies (units: 106 m2 s-1, black contours) at two transect locations: (a-h) northeastward transect location (blue solid line in Fig. 1); (i-p) northward transect location (blue dashed line in Fig. 1). The color-shaded areas are statistically significant at the 95% confidence level, based on a two-tailed Student's t-test with reduced degrees of freedom. The top-left number in each panel indicates the number of days used for each composite. The red vertical lines in (a-h) mark the locations nearest to SEP and NAH stations, while those in (i-p) mark the locations nearest to SEP and AAR stations. The values in parentheses denote degrees of latitude (positive for °N, negative for °S) and longitude (°E), respectively.
Figure7. Average ozone profiles (ozone partial pressure) in (a) BSISO phases 1-2 and 5-6 at SEP station, (b) BSISO phases 2 and 5 at AAR station, and (c) BSISO phases 8-1 and 4-5 at NAH station, from WOUDC.
Figure8. Phase points for active ISM dates (red dots) and break ISM dates (blue dots) in July-August.Phase plots are widely used in ISO studies to show the daily location of deep convection, which is then used to define the phase of each day. The normalized PC1 and PC2 of the MV-EOF analysis of OLR and zonal wind at 850 hPa are taken as the y- and x-axis of the plot, respectively, indicating the location of 30-60-day BSISO deep convection in the ISM region. Figure 8 suggests that most active ISM events occur in BSISO phases 4-7, when the 30-60-day bandpass-filtered deep convection (negative BSISO-related OLR anomalies) is over India, the Bay of Bengal, and the South China Sea (Fig. 9a). At the same time, cyclonic anomalies (Fig. 9e) and positive precipitation anomalies (Fig. 9c) are observed in these regions, causing negative ozone anomalies in the lower troposphere (Fig. 9g). Simultaneously, suppressed convection (positive BSISO-related OLR anomalies) exists over the eastern Indian Ocean (Fig. 9a). Anticyclonic anomalies (Fig. 9e) and negative precipitation anomalies (Fig. 9c) extend from the eastern Indian Ocean to the Maritime Continent, resulting in remarkable positive lower-tropospheric ozone anomalies in this area (Fig. 9g).
Figure9. BSISO-related (30-60-day bandpass-filtered) (a, b) OLR anomalies, (c, d) precipitation anomalies, (e, f) stream function anomalies (contours) and horizontal wind anomalies (vectors), and (g, h) lower-tropospheric column (700 hPa to the surface) ozone anomalies, in active (left-hand panels) and break (right-hand panels) periods of the ISM. The shaded areas are statistically significant at the 95% confidence level, based on a two-tailed Student's t-test with reduced degrees of freedom. The top-right number in each panel indicates the number of days used for each composite.On the contrary, most break events of the ISM occur in BSISO phases 8-3 (Fig. 8), when the 30-60-day bandpass-filtered deep convection (negative BSISO-related OLR anomalies) is over the eastern Indian Ocean (Fig. 9b). At this time, cyclonic anomalies (Fig. 9f) and positive precipitation anomalies (Fig. 9d) extend from the eastern Indian Ocean to the Maritime Continent, causing the prominent negative ozone anomalies in this area (Fig. 9h). In break spells of the ISM, suppressed convection (positive BSISO-related OLR anomalies) exists over India, the Bay of Bengal, and the South China Sea (Fig. 9b), accompanied by anticyclonic anomalies (Fig. 9f) and negative precipitation anomalies (Fig. 9d). In this scenario, there are positive lower-tropospheric ozone anomalies in these regions.
It is worth noting that the blue dots are farther from the center circle than the red dots in Fig. 8. This difference means that the amplitudes of BSISO are greater in break spells of the ISM than active spells, especially on some break days during phases 8-2 (Fig. 9). It suggests that BSISO conviction anomalies in break spells of the ISM are greater than in active spells (compare the amplitudes of OLR anomalies in Figs. 10a and b). As a result, BSISO-related ozone anomalies over India and the Maritime Continent seem more significant in break spells than active spells of the ISM.
To test the difference in ozone between active and break spells in the Maritime Continent, South Asia, and the western North Pacific, we also compare the ozone profiles in active and break spells of the ISM at SEP, AAR and NAH stations, based on ozonesonde observations from WOUDC (Fig. 10). At SEP station, the ozone partial pressure in active spells is greater than in break spells at almost every level under 300 hPa (Fig. 10a). This is consistent with positive ozone anomalies in active spells and negative ozone anomalies in break spells of the ISM over the Maritime Continent (Figs. 9g and h). At almost every level under about 550 hPa at AAR station, the level of ozone in break spells is greater than in active spells (Fig. 10b). This is also consistent with positive ozone anomalies in break spells and negative ozone anomaly in active spells of the ISM over South Asia (Figs. 9g and h). NAH station, located in the northwestern Pacific, where ozone anomalies possess the same BSISO variation as in the Maritime Continent (Figs. 9g and h). This is proven by the ozonesonde data at NAH station, which show that the ozone partial pressure is greater in active spells than in break spells (Fig. 10c).
Figure10. Average ozone profiles (ozone partial pressure) in active (red line) and break (blue line) periods of the ISM at (a) SEP, (b) AAR and (c) NAH ozonesonde stations, from WOUDC.Unlike wintertime MJO, which propagates eastward along the equator, BSISO shows a northeastward propagation, which means that the lower-tropospheric ozone anomalies caused by BSISO have the potentially to move further to the north to influence East Asia and the western North Pacific. The 30-60-day negative/positive ozone anomalies, accompanied by enhanced/suppressed convection, show a northwest-southeast distribution, with northeastward propagation, in the ISM region. However, there are some inconsistences between the OLR and ozone anomalies, indicating the existence of other factors that control the BSISO of lower-tropospheric ozone. Among the factors considered influential on tropospheric ozone, we find that precipitation and horizontal circulation anomalies are closely related to the BSISO of lower-tropospheric ozone. Results show that precipitation plays a dominant role in modulating the BSISO of ozone in the tropics, as it reduces precursors and temperature. The most significant negative and positive ozone anomalies occur consistently over the Maritime Continent in phases 1-3 and in phases 4-6. When BSISO propagates northeastward to the western North Pacific, horizontal transport caused by BSISO becomes relatively more important in modulating the BSISO of lower-tropospheric ozone. The southerly winds at the eastern edge of the cyclonic anomaly in the western North Pacific bring air containing low ozone in the tropics to the north, reducing the tropospheric ozone in this region in BSISO phases 7 and 8. Similarly, the positive ozone anomalies in the western North Pacific in phases 3 and 4 are caused by northerly winds at the edge of anticyclonic anomalies.
We compare the 30-60-day variations of BSISO in active and break spells of the ISM, in which the convection and rainfall in the core ISM region are extremely enhanced and suppressed, respectively. It is shown that most active ISM events occur in BSISO phases 4-7, when suppressed 30-60-day convection appears over the equatorial Indian Ocean and enhanced convection appears over India, the Bay of Bengal, and the South China Sea. Most break events occur in BSISO phases 8-3, when the convection shows an opposite pattern to that in active periods. As a result, the 30-60-day variation of tropospheric ozone shows significant positive/negative anomalies over the Maritime Continent in active/break spell of the ISM, as well as negative/positive anomalies over India, the Bay of Bengal, and the South China Sea. Also, the BSISO in break spells has larger amplitudes than in active spells, suggesting that BSISO in break spells is stronger than in active spells. Therefore, BSISO-related lower-tropospheric ozone anomalies over India and the Maritime Continent are more significant in break spells of the ISM than in active spells.
Ozone profiles observed at three ozonesonde stations from WOUDC are used in this study. The ozone at Hanoi (AAR) station represents the BSISO-related ozone variation in South Asia, while that at Sepang Airport (SEP) station represents the variation in the Maritime Continent and that at Naha (NAH) station the western North Pacific. By comparing the ozone profiles in different BSISO phases and in active/break spells of ISM, we further confirm the results regarding BSISO-related lower-tropospheric ozone from the ERA-Interim data. However, due to the limited number of stations and frequency of observation, we are unable to provide any further detail regarding BSISO-related tropospheric ozone from the WOUDC ozonesonde dataset.
Building upon previous studies of dynamic BSISO in the summer monsoon system, we focus here on the related ozone variation in the lower troposphere. As suggest in our study, BSISO in the summer monsoon region potentially modulates lower-tropospheric ozone via the precipitation and horizontal transport among the regions of mainland Asia, the western North Pacific, and the Maritime Continent. The comparison of tropospheric ozone anomalies caused by BSISO between active and break spells of the ISM warns us that BSISO-related ozone anomalies should not be ignored in the ISM region. However, the ISO of tropospheric ozone is a complicated issue, since it is influenced by emissions, transport, and other factors like air-sea interaction, which has recently been proven to be associated with ISM ISO (Zhang et al., 2018). Thus, there remains a a long way to go until we reveal the mystery of the intraseasonal variation of tropospheric ozone and other atmospheric components.
