3.1 Analysis of interannual variability of TCO in Mato Grosso do Sul
The evaluation of the behavior of the interannual and seasonal variability of the TCO for the entire MS region is presented in Fig. 1. The interannual variability of RAI+ and RAI- for the period between 2005-2020 is represented in Fig. 2 and Fig. 3. The equation of the linear regression line presents a negative angular coefficient, indicating a tendency to decrease the average values of TCO (supplementary Figure 1). It is possible to observe that the TCO values in MS in the analyzed period are typically concentrated between 254 with a minimum average of 260 and a maximum average of 258 with a standard deviation of 2.36 UD and a coefficient of variation of 5.55 DU. The existence of a semi-annual cycle of variation of the TCO was observed, having its minimum values in the April-May-June quarter of each year and the highest in the September-October-November quarters, in line with (Lima et al. 2021) with the study period of 2005 to 2015, similar to what was observed by (Lopo et al. 2013) in the period 2001 to 2009 and (Sousa et al. 2020) in the period from 1978 to 2013. The semi-annual cycle of the MS region shows the behavior of TCO in each year, the variability of the values found corroborates those obtained by (Sousa et al. 2020). In this cycle there is not great influence of the ENSO variability modes.
The variation in TCO in MS is not as pronounced as seen in (Lopo et al. 2013) and (Sousa et al. 2020), but it presents a smooth curve with a variation of approximately 6.29 DU, its minimum value in the year 2019 (254.39 UD) and the maximum value in the year 2008 (260.66 UD). In the years 2005-2011, the average TCO values remain almost stationary between 259 DU and 260 DU, in the following years it shows an intense decrease until reaching the lowest value, 254 DU in the years 2018, 2019 and 2020. This annual cycle is observed along higher latitudes, however with greater variations as shown in (Dias Nunes et al. 2020).
3.2 Analysis of Rainfall Anomaly Indices
To understand the participation of ENSO phenomena in the total annual precipitation of the series presented, the RAI adapted by (Oliveira et al. 2020) was applied. The RAI classification is performed following the values described by (Brito et al. 2022). Figure 3 shows the estimated RAI for the series of each of the eight regions, which allows classifying the years according to the proposal of (Oliveira et al. 2020) (Table 2), and thus obtain the framing of the years in Table 2.
In the analysis of the data from the historical series (2005 to 2020), for the 8 rainfall stations studied and distributed, the RAI are observed for each location. According to Reboita et al. (2010), this region has a certain climatic uniformity with regard to atmospheric mechanisms (mainly the circulation of air masses), which makes regional thermal diversification due to geographic factors such as altitude, latitude and longitude (continentality). The same author states that all static climatic factors, such as relief, act on the climate of a given region in interaction with the regional systems of atmospheric circulation, which demonstrates the importance of knowing the circulation systems that act in the region along the of the year to understand the climate.
3.3 The dynamics of the study area
Regarding the seasonal and spatial distribution of rainfall, Reboita et al. (2010) states that these are very simple, as the topography characteristics do not offer great barriers to the atmospheric circulation systems, which define the rainfall in the Brazilian Center-West Region these precipitations are not evenly distributed throughout the year. In almost all regions, more than 70% of the total precipitation accumulated during the year is from November to March, with the November-January quarter being generally the wettest. On the contrary, the winter is excessively dry and at this time of year the rains are very rare, with an average of 4 to 5 days of occurrence of this phenomenon per month, being even scarcer in the western sector of MS, where at least one month does not record a single day of rain. Therefore, drought occurs most frequently in the June-July-August winter quarter believe that only topographic factors do not play a conditioning role in the spatial distribution of these variables since the atmospheric circulation conditions are practically the same for the entire state of MS. On a macroscale, the main air masses that influence the seasonal variation and distribution of rainfall in the region are the Tropical Atlantic and Polar Atlantic (in winter) and the Tropical Atlantic Mass (in summer) – (Teodoro et al. 2016; Oliveira-Júnior et al. 2020).
The rainy season (October to March/April) concentrates over 85% of annual rainfall, with December and January contributing over 35% of annual rainfall. The rainy season (October to March/April) concentrates over 85% of annual rainfall, with December and January contributing over 35% of annual rainfall. The dry season, which begins in April and extends until the beginning of October, is characterized by a significant reduction in rainfall. In the driest quarter of the year (June-August), precipitation represents, on average, less than 2% of the annual total the daily evapotranspiration (ET) and that does not change the dryness of the environment. These periods usually exceed 100 days.
During the period of analysis, the average number of consecutive days in which prolonged droughts occurred did not exceed 75 consecutive days, with the average number of days without significant rainfall (less than 2.5 mm) being 110 days and that almost half of the years has a long period without rain of more than 75 uninterrupted days. This period coincides with the dry season, being more common in the months of June, July and August, reaching until mid-September state of MS (Teodoro et al. 2016; Oliveira-Júnior et al. 2020).
Table 2 shows that after the application of the RAI, 5.7% of the years were considered extremely humid [RAI (4)]. In turn, classified as very humid (2 ˂RAI ˂4) are 14.1% of the years and between 0 <RAI <2, that is, wet years, are 32.4%. are 5.0% of the years. In the years classified by the RAI as very dry (-4 ˂RAI ˂-2) are 17.2% and the dry RAI (-2 <RAI <0) 5.7%. Finally, the years without anomalies correspond to 3.4%. The years 2004, 2005 and 2006 (normal years) were classified by the RAI as “very dry” (-4.0 < RAI < -2.0). In 2004, 2005 and 2006, the behavior of precipitation should not be scarce, as they are under the neutrality of the ENSO phenomenon, minimum of 72 mm.
The analysis phase of the occurrence of ENSO climate variability mode was for the period 2005-2020. For this, the graphs in Figure 3 show the annual RAI+ and RAI- together with their linear trend line and the standard deviation of the series (supplementary Figure 2).
In Figure 3, it is also possible to identify years that presented precipitation within the standard deviation of the series, considered years with normal precipitation and that were under the influence of the ENSO climate variability mode, both in its positive and negative phases. The classification of the ENSO climate variability mode in years with annual precipitation totals within the standard deviation of the series: 2000; 2001; 2007; 2008; 2010 and 2011 - La Niña and the years: 2002; 2005; 2007; 2009; 2010 - El Niño (Oliveira Júnior et al. 2021).
It is possible to notice that the adoption of the standard deviation of the series as a method of verifying the influence of ENSO in the annual totals of precipitation for MS, cannot represent, with great precision, the behavior of the phenomenon, mainly in relation to the years with low annual precipitation volumes (negative phase of the phenomenon) – (Oliveira-Júnior et al. 2020).
Climatic anomalies can last several months, mainly in the tropical atmosphere, and are not only characterized by the lack or excess of some meteorological element, but also imply a change in their temporal and spatial distribution. On a global scale, the greatest influence is due to the ENSO climate variability mode and its different phases/intensities (El Niño -EN; La Niña -LN), which are closely related to changes in climate, atmospheric circulation configurations and ocean-atmosphere interaction in the Pacific and Atlantic oceans (Lyra et al. 2017), thus determining air temperature anomalies and especially rainfall in various regions (Reboita et al. 2010). ENSO can mainly influence the change in the regional rainfall regime, which can result in severe droughts or extreme rainfall, significantly interfering with human activities and alternating rainy and dry periods.
In general, the impacts of El Niño and La Niña events are known to have spatial and temporal variability, with long periods with consistent continuous anomalies not being observed at the regional scale (Gois et al. 2015). According to Souza et al. (2013) the total monthly and annual variations in precipitation are due to the behavior of the regional atmospheric circulation throughout the year, together with local or regional geographic factors, the atmospheric systems operating in Mato Grosso do Sul are: Intertropical Convergence Zone (ITCZ), Equatorial Zone Continental Tropical System (EZCTS), Atlantic Tropical System (STA), Atlantic Polar System (APS) and South Atlantic Convergence Zone (SACZ) – (Reboita et al. 2010; Oliveira Júnior et al. 2020; Oliveira Júnior et al. 2021).
The characterization of dry winters and rainy summers in Central-West Brazil stems from the stability generated by the influence of the South Atlantic Subtropical Anticyclone (SASA) and the small ridges that form over the South American (SA) continental part. The rainy season is associated with the southward shift of the ITCZ, following the apparent movement of the sun towards the Tropic of Capricorn (summer). Over the central portion of SA, the CIT advances further south than in coastal regions, generating instability throughout central Brazil in the summer months. Due to the influence of the tropical marine and equatorial air mass, temperatures are high throughout the year. In winter, when the ITCZ is shifted to the north, the region has low or no precipitation (Oliveira Júnior et al. 2021). According to Souza et al. (2013) an important climatic factor that acts in the state of MS and alters rainfall levels is altitude, which allows for differences in thermal and rainfall conditions between nearby locations (distances less than 100 km from each other).
Figure 4 shows the correlation (r) and p values for the correlation between ozone and RAI+ and RAI -, for Mato Grosso do Sul. Pearson's correlation coefficients range from -0.94 (Coxim) to -0.54 (Gold), for RAI - and ozone at 5% error probability (r ≠ 0) and for RAI+ and ozone they ranged from -0.93 (Ponta Pora) to -0.70 (Corumba).
The weak correlation indicates little influence of the El Niño and La Niña phenomena on the O3 of Mato Grosso do Sul. However, here there can often be a delay in the response of the precipitation index to the occurrence of ENSO and in the O3 column, which may, in part, explain the low correlations.
3.4 Influence of ENSO on the interannual variability of TCO in MS
The variability of TCO for the period from 2005 to 2020 in the MS region is shown in Fig. 2, where it is possible to observe the TCO values with a variation of ~2.5% in the region. The most significant variations occur in the period between August and November, a period in which there is a significant increase in TCO averages in the mid-latitudes regions due to the presence of the Antarctic Polar Vortex, which acts as a barrier preventing the arrival of the contents of ozone distributed by the CBD in the polar region, and thus, the ozone contents are trapped in the mid-latitudes region similar to what was exposed in ( Dias Nunes et al. 2020). From April to May all latitudes show a decrease in mean values.
3.5 Observed tropospheric ozone response to La Niña
Higher tropospheric ozone levels are evident at mid-latitudes, (Figure 1), mainly due to higher emission of ozone precursors Lamarque et al. (2010). To determine the relationship between MS satellite-derived tropospheric column and La Niña events, we compared the La Niña index and ozone anomaly time series. There were six La Niña and five El Niño events during 2005-2020 (see supplementary material, supplementary), implying that the irregular occurrence of ENSO does not significantly affect ozone levels. MS ozone concentration shows considerable interannual variability with a standard deviation of 2.36 DU and appears to be behind the La Nina index by a few months, implying that the mechanism of O3 variation may be associated with ENSO.
To verify the spatial distribution of O3, we evaluated the regression of tropospheric column ozone in relation to the La Niña index (supplementary material 3). A slight increase/decrease in tropospheric column ozone is observed with an average regression coefficient: R2=0.88 for Ponta Porã (RAI+) and a smaller R2=0.32 for Chapadão do Sul (RAI-) (supplementary Figures 3 and 4). Thus, ENSO can be considered the main driver of interannual variability in TCO over MS; higher and lower levels of O3 are associated with La Niña and El Niño, respectively.
As we used the linear regression method, the ozone anomalies associated with El Niño would resemble those of La Niña, but with opposite signs. In other words, El Nin leads to a decrease in tropospheric column ozone. This indicates that anthropogenic emission trends alone are not sufficient to unambiguously attribute ozone trends to global climate change, because each ENSO event has unique appearances Timmermann et al. (2018), and ENSO has been changing (Cai et al. 2018).
3.6 El Niño effects on the troposphere
ENSO also has a major influence on the interannual variability of troposphere chemistry Doherty et al. (2006). Studies of the impact of ENSO on composition (ozone and precursors) in the tropical region found an increase in tropospheric O3 column in the western Pacific and over Indonesia in El Niño years associated with dryness, downward motion and suppressed convection (Thompson 2001). High tropospheric ozone levels can be observed, coincident with fires over MS during El Niño-induced drought conditions. This increase is of interest in the impact of ENSO on tropospheric ozone, particularly in relation to atmospheric dynamics compared to the increase in biomass burning during El Niño, it can be observed that fires increase during El Niño events with an even greater increase during El Niño events extreme events. So far, the measured and simulated tropospheric ozone column anomaly patterns for El Niño have been quite consistent, showing an increase over MS. These changes in TCO are attributed to a combination of large-scale circulation processes associated with changing Walker cell convection patterns and surface/boundary layer processes due to forests in MS biomes. Changes in circulation cause a decrease in TCO associated with increased convection and upward movement of low-ozone air from the lower troposphere, and an increase associated with suppressed convection and downward movement of ozone-rich air from the upper troposphere Doherty et al. (2006). In addition to the anti-correlation between TCO anomalies and convection changes, Sudo and Takahashi (2001) found that there are other chemical changes occurring during El Niño, changes in specific humidity, cloud mass flow, and wind impact the chemical lifetime of O3. In the atmosphere.the troposphere decreased humidity results in an increase in O3 lifespan and vice-versa.
Wei et al. (2021) studied the ENSO in Asia, using satellite measurements of tropospheric column ozone and chemical-climate model simulations, found that observed ozone tends to increase in the East Asian troposphere 4 months after the La Niña peak. These post-La Niña changes are also evident in the results of the chemical-climate model, albeit with a slightly longer delay (5 months).
Lima et al. (2021) analyzed the annual variation and effects of the El Niño atmospheric variability mode (Canonical and Modoki) on the TCO over Northeast Brazil (NEB) between 1997 and 2018 using data from Total Ozone Mapping Sensors Spectrometer (TOMS) and OMI. There was an average monthly variation throughout the studied area with typical behavior of the annual cycle in seasonal variability, with a minimum value in May and a maximum in October. A downward trend was observed in the series of mean values during the analyzed period. The averages of the anomalies show that El Niño events affect the TCO, predominantly causing a decrease in its values. These events in Modoki mode have greater potential to affect the TCO than Canonicals with negative anomalies of greater intensity.
Yang et al. (2022) studied different impacts of hot/cold phases of ENSO on summer tropospheric O3 over China based on model simulations, soil measurements and reanalysis data. Summer surface O3 concentrations in China show a positive correlation with the ENSO index during the years 1990-2019, with the largest increases of 20% in southern China in El Niño (warm phase) versus La Niña years (cold phase). Furthermore, the increase in O3 during El Niño years is mainly due to domestic emissions in China. This study highlights the potential significance of ENSO in modulating tropospheric O3 concentrations in China, with major implications for the mitigation of O3 pollution.