*3.1. Unusual Behaviour of Trace Gases in the Tropical UTLS Region during the 2015–16 El Niño Event*

The monthly mean time series of observed O3 at 100 and 82 hPa and CO at 146 and 100 hPa over the equatorial region (averaged over 10◦N and 10◦S) along with MEI index from January 2006 to December 2017 are shown in Figure 1. O3 shows significant seasonal variations with high values during Northern Hemisphere (NH) summer months and low values during NH winter months (Figure 1a). However, during the year 2015 a drastic change in O3 mixing ratio at both pressure levels is noticed particularly during the summer months of July and August in 2015 as compared to the other years. The MEI index also shows the gradual increase in its strength and reaches record high values (>2 MEI index value) in September 2015 (Figure 1c). After that, the MEI index maintains its value of ~2 till May 2016. It is well known that the ENSO dominates the inter-annual variability in the O3 at tropopause level [10,32]. The CO (Figure 1b) also shows a strong enhancement in September, October and November 2015. The values are very high during 2015 as compared to the other years. The O3 and CO show significant decrease and increase. The relative percentage changes in the O3 and CO with respect to the averaged period from 2006 to 2014 are presented in the Section 4, respectively.

The monthly climatological (2006–2014) means of O3 (CO) mixing ratio at 146, 100 and 82 hPa (146, 100 and 68 hPa) in the tropics (20◦N–20◦S) along with the seasonal change in the year 2015 are illustrated in Figure 2. Three latitude bands within the tropical latitudes, namely equator (averaged over 10◦N–10◦S), Northern Hemisphere (NH) (averaged over 11◦N–20◦N) and Southern Hemisphere (SH) (averaged over 11◦S–20◦S) are selected to obtain the variability in the O3 mixing ratio. We can see the clear seasonal variability in the climatological O3 mixing ratio at both the pressure levels over the tropics.

**Figure 1.** Time series of observed monthly mean (**a**) ozone at 100 and 82 hPa, (**b**) carbon monoxide at 100 and 146 hPa and (**c**) Multivariate ENSO Index (MEI) from January 2006 to December 2017.

**Figure 2.** Annual cycle of ozone mixing ratio observed at 146 hPa (**a**–**c**), 100 hPa (**d**–**f**) and at 82 hPa (**g**–**i**) averaged over different latitude bands. Black colour line shows monthly climatology of ozone mixing ratio calculated by using MLS data from January 2006 to December 2017 and red colour line shows the monthly mean of ozone mixing ratio during 2015. Vertical bars indicate standard deviations of the measurements.

Note that over the equator (Figure 2e,h) and SH (Figure 2f), the O3 shows a quite different seasonal change in the year 2015 compared to the climatology. A clear drop in the O3 mixing ratio in the equator from June to September is observed (Figure 2e,f). In 2015, from June to December, the seasonal change is completely disappeared in the tropical regions at 100 hPa and 82 hPa as seen in Figure 2e,h, respectively. In NH, the climatological and the year 2015 O3 values show a clear distinct picture. At 100 (82) hPa, the O3 value from January to May follows the climatological pattern and then in June it starts to deviate from it. The difference is greater at 100 hPa as compared to 82 hPa. From these results, it is clear that O3 shows a significant drop in the tropics from June to September 2015. Similar unusual changes are noticed in the CO mixing ratio over the tropics during this period but a little bit later (Figure 3). Figure 3 shows that, in the entire tropics, CO shows a strong enhancement in the months from September to December at both pressure levels (146 and 100 hPa) as compared to the climatology. The monthly mean values of O3 and CO over the tropics clearly indicate that strong and unusual changes occurred in the UTLS region during the 2015–16 El Niño event.

To quantify the changes in the trace gases during the year 2015, we have estimated the monthly anomalies by subtracting the climatology from the individual monthly mean trace gas concentrations. The time series of O3 and CO anomalies observed at different levels in the UTLS region over the 10◦N–10◦S region is shown in Figure 4. Strong negative anomalies in O3 are observed at 82 and 100 hPa but not at 146 hPa (Figure 4a). The O3 decrease is high at 82 hPa (~80 ppbv) as compared to 100 hPa (40 ppbv). This decrease is very high from September to December 2015 (peaking in November) with large negative anomalies and continued in 2016 with less magnitude. This might be due to the strengthening of the Brewer Dobson circulation (BDC), that is, strong enhancement of tropical upwelling during El Niño period [9–11]. Previous studies clearly demonstrated the impact of the ENSO on the inter-annual variability of O3 at the tropopause level [32]. These inter-annual variations in the O3 anomalies are linked with the SSTs in the equatorial Pacific Ocean and are explained by the strength of the BDC. In general, during the El Niño period, the tropical upwelling increases whereas in La Niña events, the tropical upwelling decreases. Note that previous studies focused on the O3 and

other trace gases changes mainly in the boreal winter (mainly in December when the intensity of the El Niño becomes peak). However, the observed high decrease of O3 in the present study is in July and August that belong to boreal summer. These results matched well with those reported by Tweedy et al. [17] and Diallo et al. [11]. They clearly demonstrated that the changes in the O3 anomalies in NH are due to the meridional advection in northern subtropics altered by boreal summer ENSO events and in SH due to the tropical upwelling [17]. The structural changes in the BDC due to 2015–16 El Niño event is also one of the reasons for the O3 anomalies observed in the tropical LS during this event [11]. Overall, instead of chemical reactions, the transport processes (due to BDC) are the major possible reason for the presently observed O3 anomalies. The CO anomalies (20–35 ppbv) show a substantial increase in the UTLS (mainly 146 and 100 hPa) from September to December 2015 (peaking in November) as compared to the whole MLS data period (Figure 4b). Note that the observed CO anomalies in 2015 are a record high in the 21st century. From Figure 4, it is evident that the O3 drop was high from June to September 2015 (peaking in August) while strong enhancement in CO was observed from September to December 2015 (peaking in November). More significantly, the maximum CO anomalies are noticed in the troposphere in October whereas in the LS (~100 and 68 hPa), it was in November (Figures 7 and 11). In the following section, we investigate the spatial distributions of these anomalies in 2015.

**Figure 3.** Same as Figure 2 but for the monthly mean of carbon monoxide at 146 hPa 100 hPa and 68 hPa.

**Figure 4.** Time series of de-seasonalized anomalies of (**a**) ozone and (**b**) carbon monoxide observed over Equatorial region (10◦N and 10◦S) from January 2006 to December 2017. Different colours indicate different pressure levels.
