Tropospheric ozone (O3) has a significant impact on the oxidative capacity of the troposphere by being a major source of hydroxyl radicals (OH) (e.g., Lelieveld et al., 2016; Logan et al., 1981) and also on climate by being a powerful greenhouse gas (e.g., Myhre et al., 2013). The first source of tropospheric ozone is photolysis of NO2, mainly formed by photochemical cycles involving NOx, HOx, and volatile organic compounds (VOCs) or carbon monoxide (CO) (e.g., Seinfeld and Pandis, 2016). O3 is a secondary compound with a shorter average lifetime (19 d, Murray et al., 2014) whose contributions are difficult to identify. Indeed, O3 concentrations in the free to upper troposphere (UT) are influenced by (i) transport from the stratosphere (e.g., Hsu and Prather, 2009; Stevenson et al., 2013); (ii) tropospheric transport (e.g., Barret et al., 2016; Sauvage et al., 2005, 2007b; Zhang et al., 2016); (iii) emissions of precursors (VOCs, CO, HO2); (iv) NOx emissions by surface combustions but also by free-tropospheric lightning (e.g., Barret et al., 2010; Sauvage et al., 2007c); or by (v) the elimination of O3 by reaction with HO2, OH or photolysis in particular (e.g., Crutzen et al., 1999). Due to its 2-month lifetime (e.g., Edwards et al., 2004), CO is a good tracer of combustion processes (Granier et al., 2011) which emit O3 precursors (i.e., NOx, VOCs, CO). CO impacts the oxidative capacity of the troposphere by being a major sink of OH radicals in non-polluted atmosphere (e.g., Lelieveld et al., 2016) and therefore the climate through enhancement of the lifetime of CH4, the second most important greenhouse gas (GHG). Oxidation of CO produces CO2 and ozone, which are also GHGs (e.g., Myhre et al., 2013). If it is partly primary (directly emitted into the atmosphere from anthropogenic combustion or forest fires), CO can also be secondary by being formed during the oxidation of VOCs.

The intertropical troposphere is a region of great interest regarding the photochemistry and energy budget of the atmosphere. It indeed combines (i) strong photochemical activity due to intense solar radiation; (ii) high emissions by biomass fires, vegetation and lightning; and (iii) dynamic processes that allow the redistribution of chemical species at regional and global scales. It is therefore an area favorable to the formation of O3. It is also one of the areas of the world where the increase in O3 mixing ratio has been most pronounced since 1980 (Zhang et al., 2016). Africa has the largest continental surface of the intertropical zone and is also a region where high CO mixing ratios (locally above 150 ppb around 250 hPa, e.g., Barret et al., 2010; Cohen et al., 2018; Fu et al., 2016; Sauvage et al., 2007d) are found in the UT. Circulation patterns above intertropical Africa impact the distribution of chemical compounds in the area (e.g., Sauvage et al., 2005, 2007a, c). One of the main circulation patterns is formed by winds from subtropical latitudes to the Equator. They converge at the Intertropical Convergence Zone (ITCZ) before being redistributed vertically, thus forming the Hadley cells. The meridional position of the ITCZ has a strong seasonal variability over Africa (e.g., Suzuki, 2011), causing the alternation of dry and wet seasons. This meteorology particularly impacts the O3 and CO distributions in the African UT (e.g., Sauvage et al., 2007d) but also their propagation in the global atmosphere (e.g., Zhang et al., 2016).

Despite the importance of this area, few observations have been carried out and Africa remains little documented in comparison with the northern mid-latitudes. The specific meteorology of the zone makes it difficult for the models to reproduce the concentrations of O3 and CO, notably at high altitude (e.g., Barret et al., 2010; Sauvage et al., 2007a). Satellite observations have been used to document the composition of the African UT, but they have a rather coarse vertical resolution (Barret et al., 2008, 2010).

Since 1994, in the framework of the In-service Aircraft for a Global Observing System program (IAGOS, previously named MOZAIC; Marenco et al., 1998; Petzold et al., 2015; http://www.iagos.org, last access: 20 September 2020), airborne measurements of chemical compounds have been made over Africa. Using upper-tropospheric measurements of O3 and relative humidity (RH) from 1994 to 2004, Sauvage et al. (2007d) showed an ozone minimum collocated with a RH maximum at the ITCZ. Both vertical transport in the ITCZ followed by photochemical production in the upper branches of the Hadley cells were found to contribute to meridional O3 gradient creation. Few CO measurements were available at this time and none south of the ITCZ.

Between December 2005 and 2013 (2006–2013 hereafter), there were almost daily flights by Air Namibia equipped with IAGOS instruments between Europe and southern Africa, thus providing a unique and highly representative dataset of in situ O3 and CO on both sides of the ITCZ. The SOFT-IO model (Sauvage et al., 2017b) has been developed to complement the IAGOS database to estimate anthropogenic and fire contributions to CO for each IAGOS measurement. These measurements, together with the SOFT-IO model outputs, allow us to trace the distributions, variability, and origins of O3 and CO in the African UT. Since 2008, the Infrared Atmospheric Sounding Interferometer (IASI) instrument aboard the Metop-A satellite has also documented UT CO and O3 global distributions (Barret et al., 2016; Tocquer et al., 2015) and therefore complements the African transects provided by IAGOS.

The objectives of this study are to (i) analyze the meridional distribution of O3 and CO mixing ratios observed by IAGOS between 2006 and 2013, (ii) explore sources and geographical origins of observed CO anomalies (calculated with the SOFT-IO model), and (iii) give first track on which transport processes drive the CO and O3 meridional transects observed especially using retro plumes (calculated with the Lagrangian particle dispersion model FLEXPART) of air masses measured by IAGOS. Observational (IAGOS, IASI) and model-based (SOFT-IO) datasets and methods are introduced in Sect. 2. In Sect. 3, the method to locate the ITCZ and establish the seasonality with IAGOS observations is described and validated. In Sect. 4, O3 and CO seasonal meridional transects observed by IAGOS are analyzed and compared to IASI observations. In Sect. 5, origins and sources of observed CO are explored with SOFT-IO, and, finally, characterization of O3 transects is discussed in Sect. 6.

Ozone precursors are emitted in large quantities across the African continent (Liousse et al., 2014; Sauvage et al., 2007b), but our understanding of the transport and effects of these precursors is limited by a lack of in situ data in the African UT. The climate of southern and northern Africa is characterized by alternating wet and dry periods which are modified by the position of the ITCZ. Within the ITCZ warm and humid surface air masses converge and are convectively uplifted into the upper troposphere. The position of the convergence zone gives rise to the “wet” seasons. The uplifted air masses are then advected polewards in the upper branches of the Hadley cells. In the upper troposphere, the air masses are zonally transported by (i) the northern and southern subtropical jets, (ii) the flow from the South Atlantic anticyclone, and (iii) the easterly winds of the tropical easterly jet (TEJ) and Asian monsoon Anticyclone (AMA) (in boreal summer) (e.g Krishnamurti et al., 2013). The dry air in the descending branches of these cells creates the conditions for wildfires and the resulting emission of ozone precursors. Seasonality in Africa is thus closely related to the geographical shift of the ITCZ from the Northern Hemisphere during the boreal summer to the Southern Hemisphere during the boreal winter. The classic four seasons of 3 months are therefore not adapted to our vast area of study centered on the African inter-tropical zone, and it was therefore necessary to establish a new definition of seasons.

Figure 6 represents, averaged between 2006 and 2013, IAGOS CO for the four seasons (Fig. 6a–d) and the contributions to these CO mixing ratios from anthropogenic and wildfire sources (Fig. 6e–h) as estimated with SOFT-IO. These anthropogenic and wildfire sources are further decomposed by region in Fig. 6i–l and m–p respectively. The different regions are denoted by their GFED acronym (see Fig. 1). It is worth recalling that the background is not simulated by SOFT-IO (see Sect. 2.2.1).

The meridional transects of total CO anomalies from SOFT-IO exhibit similar behavior to the IAGOS data with CO mixing ratios characterized by a clear maximum shifting from the north of the ITCZ in DJFM to the south of the ITCZ in JJASO. In DJFM, the fire and anthropogenic contributions shown by SOFT-IO are in phase, with the maximum CO, as with the IAGOS observations, found at the Equator. In JJASO, the maximum contribution from fires and the maximum observed CO both occur at 5∘ S, while the maximum contribution from anthropogenic sources is located at 12∘ N. The resulting total contribution has therefore shifted north (∼Equator) relative to the observed one. This may indicate an underestimation of SHAF fire and/or anthropogenic contributions or an overestimation of the anthropogenic contribution from the southeast and central Asia (SEAS and CEAS) regions which are responsible of the high CO north of the Equator (see Fig. 6k).

The absolute quantitative CO mixing ratios estimated with SOFT-IO are subject to biases and uncertainties especially due to uncertainties in emission inventories and difficulties in representing tropical deep convection either by the ECMWF analyses or by the FLEXPART model. However, the comparisons between the different estimated contributions remain consistent. The anthropogenic and fire CO sources are of the same order of magnitude during the main seasons, while the anthropogenic source dominates during the transition seasons due to a decrease in fire contributions. Overall, the contribution of fires is not the dominant contribution to CO in the African upper troposphere, consistent with the observed and continuous increase in anthropogenic emissions (e.g., Liousse et al., 2014) and the slight decrease in fire emissions (GFAS, Kaiser et al., 2012).

The anthropogenic contributions are mainly local, with Northern Hemisphere and Southern Hemisphere Africa regions of GFED (NHAF and SHAF; see Fig. 1) representing about 66 % to 90 % of the estimated anthropogenic contributions at the location of the CO maxima. The local contributions have rather low seasonal variabilities. As expected SHAF contributes mostly south of the Equator (maximum between 5 and 10∘ S) and NHAF north of the Equator (maximum around 10∘ N). The CO emitted locally at the surface is transported in the lower troposphere by the trade winds towards the ITCZ and then uplifted to UT (Edwards et al., 2003; Sauvage et al., 2006). Thus, in DJFM (JJASO), the maximum contribution of SHAF (NHAF) is co-located with the position of the ITCZ. In addition to local influence, anthropogenic CO in the African UT is influenced by long-range transport from Asia. The contribution from southeast and central Asia (SEAS and CEAS) has an important seasonal variability. The SEAS anthropogenic CO contribution is negligible in the period April–May but represents up to 30 % of total CO anthropogenic contribution in DJFM and 21 % in November at the maximum located at the Equator. The Asian contribution is the strongest in JJASO in the Northern Hemisphere, representing up to 66 % of estimated anthropogenic contributions at 20∘ N. Barret et al. (2008) have shown that this strong Asian contribution results from South Asian emissions uplifted to the UT by the deep convection associated with the monsoon and then transported westward by the tropical easterly jet. We also observe an important contribution from the Middle East (MIDE) in JJASO, centered at the ITCZ. This MIDE contribution corresponds to surface air masses from North Africa and Arabia, which followed the same circulation as local surface air masses to UT ITCZ. In the Northern Hemisphere, a small contribution from temperate and central America regions (TENA and CEAM) is observable during all the seasons, transported via upper-level westerly winds close to the subtropical jet.

The fire contributions are predominantly local (NHAF and SHAF) for all the seasons with an important seasonal variability. The geographical origin of CO from fires follows the dry and wet seasons of the two hemispheres. During the dry season of the Northern (Southern) Hemisphere, NHAF (SHAF) fires contribute more than 90 % to the meridional CO transects. The maximum contributions (≈13–14 ppb, i.e., in addition to the background) are shifted by approximately 10∘ towards the dry zone relative to the ITCZ. This is explained by the location of the emission zones (i.e., dry zones), which are shifted relative to the ITCZ. CO emitted at the surface is transported at low altitudes towards the ITCZ by the northeast trade winds and the Harmattan in DJFM and by the southeast trade winds and the South Atlantic anticyclone in JJASO (Edwards et al., 2003; Sauvage et al., 2006). The air masses are then uplifted by deep convection and move away from the ITCZ in the upper branches of the Hadley cells (e.g., Barret et al., 2010; Sauvage et al., 2007c). The contribution to the CO from fires in the transition periods (maxima ≈5 ppb in April–May and ≈9 ppb in November above background levels) is mostly from SHAF in the Southern Hemisphere and from NHAF in the Northern Hemisphere. The contributions from fires are lower during these transition periods due to the moisture retained from the previous wet season or the increasing humidification going into the next wet season. Small contributions from Indonesia (EQAS) are observed in JJASO (maximum relative contribution ≈10 % at 20∘ N) and in November (≈20 % at 0∘ N). They correspond to air masses advected by the TEJ and diffused north and south during the transport. South American (SHSA) contributions are also observed in JJASO in the Southern Hemisphere (maximum relative contribution ≈30 % at 20∘ S) and in November on the entire transect (maximum relative contribution ≈50 % at 20∘ S). The polluted air masses are transported in the UT by the westerly winds of the South Atlantic anticyclone in the Southern Hemisphere and by the subtropical jet in the Northern Hemisphere.

It is interesting to note that, in both DJFM and JJASO, the CO plumes are transported over the Gulf of Guinea by the easterly winds that dominate within the ITCZ (see Fig. 3) as documented by IASI (Fig. 5). The IASI distributions also highlight that during JJASO the strong TEJ (Fig. 3c) transports the CO plume westwards while during DJFM the transport is north-westward due to a northwards wind component over Africa (Fig. 3a).

In the previous section, we used the SOFT-IO tool to attribute high concentrations of CO to anthropogenic and wildfire sources across different regions. As tropospheric O3 is a secondary pollutant, the source attribution is more complex, as the ozone may be lost or produced along its trajectory from the surface. There may be local sources in the upper troposphere such as intrusions of ozone-rich air from the stratosphere and photochemical production from NOx generated by lightning (LiNOX), or even via NOx emitted by aircraft. The analysis of the meridional transects of ozone is more complex. It is possible to determine the origin of air masses and the spatial variability of ozone and give the associated transport pathways, independently of the CO mixing ratio. Therefore, in this section we do not use the SOFT-IO as we did for CO, but we look at the FLEXPART backward plumes initiated at the IAGOS data points. Based on previous modeling studies (Sauvage et al., 2007a–c), we hypothesize that O3 is produced during the transport of air masses impacted by precursors emitted at the surface (such as CO as described in the previous section) and sources at altitude such as lightning as inferred in the Sauvage et al. (2007b) study with a chemistry and transport model.

Figures 7 and 8 represent the FLEXPART backward plumes corresponding to air masses characterized by O3 maxima and minima as observed by IAGOS in DJFM (Fig. 7) and JJASO (Fig. 8). The backward plumes are displayed as the integration of 20 d residence times (equivalent to the origin probability density; see Stohl et al., 2005) summed vertically and averaged over the whole period (end 2005–2013). For each case, backward plumes are computed for sections of 5∘ of latitude from the corresponding IAGOS flights. The highest altitude level corresponds to an 11 km and above level which can include both UT and LS. Determining the altitude of the intertropical tropopause with the potential vorticity to distinguish tropospheric or stratospheric origins is complicated. As the contributions in this layer do not extend beyond the tropics (±25∘ approximately) where stratospheric incursions are negligible, the approximation of a predominantly UT origin of these contributions is made.

In DJFM, the O3 minimum is observed at the ITCZ (≈11∘ S; see Fig. 4) and corresponds to African air masses (Fig. 7e) rising from the surface (Fig. 7d and f). O3 mixing ratios are globally lower in surface air masses due to either clean maritime air masses or O3 titration in fresh polluted air masses. The area of origin of the air masses (Fig. 7.e) corresponds to the convergence zone delimited by the high RH and the OLR below 220 Wm-2 in Fig. 3e. We note that the ascension of the surface air masses takes place to the southeast of IAGOS aircraft, i.e., in the center of this convergence zone, where the Hadley cell vertical wind speed is the strongest.

The maximum of O3 observed by IAGOS in DJFM (≈10∘ N; see Fig. 4) corresponds to high-altitude air masses transported during the previous 20 d (Fig. 7a–c). These are (i)

air masses uplifted at the ITCZ and transported away from the ITCZ by the Hadley cell upper branches – chemical processing during transport allowed O3 production from African LiNOX emitted in the UT at the ITCZ and from uplifted African surface emissions (CO maxima at ≈2∘ N);

As already discussed (Sect. 4.2), in DJFM there are two O3 maxima documented by IASI on each side of the ITCZ which stem from the same crescent-shaped plume (Fig. 5c). The northern African O3 maximum is located a bit to the south relative to the CO maximum according to IASI (Figs. 5a and c and 4a and c) and a bit further north according to IAGOS (Fig. 4a and c). Nevertheless, in both cases they are located on the same side of the ITCZ. The FLEXPART run indicates that at the location of the northern O3 maximum seen by IAGOS, air masses are mostly coming from the West African fire region explaining the coincidence between the O3 and CO plumes. The air masses observed by IAGOS at 20∘ S (local maximum of O3) may have been thus transported in the UT from the Gulf of Guinea to the South Atlantic Ocean and then to the south of Africa where they have been sampled by IAGOS. Mixing with surrounding O3-poorer air masses during this transportation is probably compensated for by the input and photochemistry of precursors from South America and lightning.

In JJASO, the minimum of ozone observed by IAGOS at the latitude of the ITCZ (≈6∘ N; see Fig. 4) corresponds mainly to O3-poor African air masses (Fig. 8e) convectively uplifted from the surface (Fig. 8d and f) and transported westward during the convection. Clean air masses from the Indian Ocean region convectively uplifted and transported by the TEJ also contribute to the O3 minimum. During this season the O3 maximum (≈15∘ S; see Fig. 4) corresponds to air masses which have been uplifted at the ITCZ in Africa (northeast of the aircraft position) or in South America more than 20 d before the observations. (Fig. 8a and c). The coincidence between the CO and O3 plumes from IASI (Fig. 5b and d) indicates that the O3 production is directly impacted by African emissions.

North of 20∘ N, a decrease in O3 mixing ratios in DJFM is observed, which is not observed in JJASO (see Fig. 4). To understand this difference, the FLEXPART backward plumes corresponding to IAGOS flights between 20 and 25∘ N for the two seasons are presented in Fig. 9. In the two cases, the air masses mainly resided above 9000 m in the 20 d preceding the measurements. In DJFM (Fig. 9a–c), the low O3 mixing ratios correspond to high-altitude old and clean air masses transported by the subtropical jet up to 20 d before the measurements. Some of the air masses come from the Atlantic lower troposphere north of the ITCZ between the Equator and 10∘ N more than 20 d before the measurements. Therefore, these air masses have a clean origin, which explains their low O3 content. In contrast, in JJASO (Fig. 9d–f), the air masses measured at around 25∘ N come from both east and west. Part of the air masses have been convectively uplifted from the South and South East Asian highly polluted boundary layer before being transported over northern Africa at the southern flank of the AMA (Barret et al., 2008). The difference resulting from the uplift of these polluted air masses and from the uplift of cleaner air masses from the Indian Ocean more to the south is highlighted by the IASI CO UT distribution characterized by a clear latitudinal gradient over the Arabian Sea and India (Fig. 5b). The chemical aging of these polluted air masses during their transport results in high O3 levels in the North African UT than in DJFM. This important contribution of precursors from Asian anthropogenic sources in JJASO is also highlighted by the SOFT-IO data presented in Fig. 5k.

It is also interesting to note that the ozone maxima observed beyond the dry zones (by ≈10∘ N in DJFM and by ≈15∘ S in JJASO) are higher in JJASO than in DJFM, while CO levels are quite similar (see Fig. 4). This probably indicates other sources of precursors not correlated with CO in the southern part. Indeed, previous studies have highlighted a zone of high O3 mixing ratios over the southern Atlantic during the boreal summer: the zonal wave one (Sauvage et al., 2006; Thompson et al., 2003) which results from (i) high precursor emissions in Africa and South America, (ii) strong solar radiations, (iii) the high thunderstorm activity and LiNOx production at the ITCZ, (iv) stratospheric O3 incursions, and (v) wind currents from the South Atlantic anticyclone.

Near-daily IAGOS airborne measurements carried out from the end 2005 to 2013 between Europe and Namibia were used to analyze and characterize meridional CO and O3 transects in the African upper troposphere and their seasonal variability. First, we developed a methodology to localize the ITCZ with IAGOS meteorological data (wind and over-water RH data) which allowed us to delimit empirically two periods of the year of strongest interest in this region, namely June to October (JJASO) and December to March (DJFM), with the intermediate months corresponding to transitional periods. The ITCZ was found to extend over 14.5–4.5∘ S during JJASO and 4.5–9.5∘ N during DJFM, in good agreement with NCEP/NCAR reanalysis.

For the two main seasons the CO transects are characterized by peak distributions with maximum mixing ratios located 10∘ from the position of the ITCZ above the dry regions (132 to 165 ppb at 0–5∘ N in DJFM and 128 to 149 ppb at 3–7∘ S in JJASO). Maximum CO mixing ratios display strong interannual variability with up to 30 ppb between years over the end of the 2005–2013 period. The O3 transects are characterized by mixing ratio minima of ∼42–54 ppb at the ITCZ (10–16∘ S in DJFM and 5–8∘ N in JJASO) framed by local maxima (∼53–71 ppb) coincident with the wind shear zones north and south of the ITCZ. The O3 transects display no clear interannual variability in terms of gradients and location of the extrema but large mixing ratio variability (∼15 ppb).

CO and O3 UT African transects are also computed from IASI 2008–2013 retrievals. Comparisons show that IASI is able to capture the main features of the CO and O3 distributions in the African UT. The CO maxima are coincident with IAGOS but the average CO mixing ratios from IASI are biased low by about 20 ppb. Concerning O3, the IASI mixing ratios are in good agreement with IAGOS, but the latitudes of the extrema can be shifted by up to 10∘ relative to IAGOS. The IASI 2D CO and O3 distributions have therefore been very useful to complement the 1D IAGOS meridional transects. IASI UT O3 distributions in DJFM have in particular revealed a crescent-shaped O3 plume above the Atlantic Ocean around the Gulf of Guinea.

SOFT-IO simulations first showed that anthropogenic and fire sources of CO have an impact of the same order of magnitude in the African UT. In DJFM both contributions are coincident and both sources drive the maximum of the CO mixing ratio. In JJASO, the observed CO maximum is coincident with the fire contribution. The important Asian anthropogenic contribution leads to a northward-shifted total contribution maximum not coincident with the observed one. This tends to indicate either an overestimation of the Asian anthropogenic contribution or an underestimation of uplifted local contributions to the African UT CO maximum in JJASO.

The anthropogenic CO contribution is mostly local (northern and southern Africa) with a low seasonal variability. CO emitted locally at the surface is advected by the trade winds towards the ITCZ where it is convectively uplifted. The large Asian contribution is related to the fast convective uplift of polluted air masses in the Asian monsoon region which are further westward transported by TEJ and AMA. The fire contribution is almost entirely local, with a significant seasonal variability following the alternation of dry and wet seasons.

For the two main seasons, the CO African UT transects characterized by maxima over the dry regions result from the following processes. CO emitted by fires in the dry region is transported towards the ITCZ by the trade winds and further convectively uplifted. In the UT the CO-enriched air masses are transported away from the ITCZ by the upper branches of the Hadley cells and accumulate within the zonal wind shear zones where the maxima are located.

The FLEXPART backward plumes revealed that (i) the O3 minima correspond to air masses that were recently uplifted (i.e., for less than 20 d) from the surface where mixing ratios are low at the ITCZ, (ii) the O3 maxima correspond to old high-altitude air masses (i.e., uplifted from the surface for more than 20 d before the IAGOS measurement) either locally transported from Africa in the upper branches of the Hadley cells or transported over long distances from South America by the west winds of the South Atlantic anticyclone and subtropical jets, and (iii) the high O3 concentrations in JJASO over North Africa around 25∘ N are related to pollution recently uplifted in the Asian monsoon region and transported south of the AMA.

All these air masses came from areas where O3 precursor emissions are high (Africa, South America and South Asia). We thus do the hypothesis of the O3 formation in these air masses during the transport in the UT due to the combination of high concentrations of precursors (from lightning at ITCZ and surface emissions) with strong radiations enhancing the photochemistry. The more time the air mass spends in the UT, the more O3 is formed (through photochemical production). The comparison of the origin of the air masses recorded by IAGOS flights in the Northern Hemisphere shows that, in DJFM, without the Asian contribution, the O3 mixing ratios start to decrease. It highlights the importance of the precursor load of air masses to fuel the photochemistry in the UT. In the Southern Hemisphere, the IAGOS O3 data allow us to further characterize the zonal wave one phenomenon with similar CO concentrations in DJFM and JJASO but higher O3 concentration in JJASO.

This analysis of meridional transects contributes to a better understanding of distributions of CO and O3 in the intertropical African upper troposphere and the processes which drive these distributions. Therefore, it provides a solid basis for comparison and improvement of models and satellite products in order to get the good O3 for the good reasons.

However, the study only covers the 2005–2013 period, and the conclusions could evolve notably according to the projections of future emissions of precursors. Indeed, the analysis of data from the emission inventory MACCity indicates that African anthropogenic emissions of CO have increased by +283 % for NHAF and +256 % for SHAF between 1960 and 2020 (against +63 % for the entire world). CO emissions from these two African regions accounted for 7.87 % of global emissions in 1960 compared to 17.96 % in 2020. On a smaller timescale (end 2005–2013) this represents a +12 % increase in CO emissions for NHAF and +16 % for SHAF (+1.6 % for the entire world). During the same period the part of both regions in anthropogenic global CO emissions goes from 15 % to 16.8 %. If African anthropogenic emissions continue to increase as fast as estimated, they could modify the typical distribution of CO and thus of O3 with first higher concentrations of CO to fuel O3 formation during all seasons.