Divergent features of the upper-tropospheric carbonaceous aerosol layer: effects of atmospheric dynamics and pollution emissions in Asia, South America, and Africa

The ozone mapping and profiler suite (OMPS) limb profiler (LP) aboard the Suomi National Polar-orbiting Partnership satellite was launched in October 2011, aiming to retrieve vertical profiles of ozone, aerosol extinction, and cloud-top height (Chen et al 2018, 2020). The instrument measures limb-scattering radiance at six wavelengths (510, 600, 675, 745, 869, and 997 nm) in the 0–80 km altitude range. The relatively high vertical and spatial sampling allow detection and tracking of sporadic events when aerosol particles are injected into the tropopause (Wu et al 2022a). In studying the TCAL, this work used the Level-2 swath observation product of the OMPS/LP (Taha et al 2021). We note that the OMPS/LP retrieved the signals of total aerosols components, which was only utilized to verify the existences of this phenomenon.

In addition, aerosol optical depth (AOD) from CALIPSO (Winker et al 2010), land cover and fire monitoring product from the moderate resolution imaging spectroradiometer (MODIS) instrument (Justice et al 2002), outgoing longwave radiation (OLR) from US National Oceanic and Atmospheric Administration (NOAA) satellite observations (Gruber and Krueger 1984), and global precipitation climatology project (GPCP) combined the satellite retrieval and rain gauge observation (Huffman et al 1997) were also applied.

The modern era retrospective analysis for research and applications version 2 (MERRA2) atmospheric reanalysis generated by the US NASA global Modeling and Assimilation Office provides traditional atmospheric products and chemical compositions for aerosols and greenhouse gases (Buchard et al 2017, Gelaro et al 2017). MERRA2 modeling and assimilation have allowed many advances in the understanding of atmospheric chemistry, ozone, and stratosphere processes (Randles et al 2017). Aerosol and ozone products have proved very helpful in the study of global air pollution, from the troposphere to the stratosphere (Che et al 2019, Wang et al 2021). Moreover, the tropopause aerosol layer phenomenon has been revealed by several studies utilizing MERRA2 reanalysis (Yuan et al 2019a, Lau and Kim 2022, Wu et al 2022b, Bresciani et al 2024). The MERRA2 has been employed to recognize CAs trend in the upper troposphere and lower stratosphere linked to the anticyclone, and the aerosol components including CAs, dust and sulfate have also been widely utilized to explore the formation, source, transport, and variability of tropopause aerosol layer with the validation of observation including CALIPSO and in-situ measurements (Yuan et al 2019a, Lau and Kim 2022, Wu et al 2022b). The vertical profiles of aerosol from MERRA2 have also been validated against CALIPSO satellite retrieval and in-situ observation (Mahnke et al 2021) as shown in figure S1. Therefore, the aerosol and meteorology-related products were thus used here to explore atmospheric dynamics influences on the TCAL.

The fifth generation European re-analysis (ERA5) was developed by the European Centre for Medium-Range Weather Forecasts to model data, physics, and core dynamics to provide a detailed description of the global atmosphere, land surface, and ocean waves (Hersbach et al 2020). ERA5 meteorological data were also used in this study.

Models of the sixth phase of the coupled model inter-comparison project (CMIP6) enable long-term simulations and various experiments in reconstructing the historical evolution and future projection of climate change and air pollution (Eyring et al 2016, Bauer et al 2020, Zanis et al 2020). CMPI6 model outputs were used to verify the occurrence of TCALs.

Weighted multi-model ensemble means (WMMEMs) were used in model bias correction to reduce errors associated with the inter-model spread of CMIP6 data as follows (Shen et al 2021):

$$\begin{equation}Y\left( t \right) = \,\mathop \sum \limits_i {W_i}{X_i}\end{equation} \tag{ 1 }$$

where ${X_i}$ is the output time series of each model, ${W_i}$ is the weighting of each model error relative to a MERRA2 reference, and $Y\left( t \right)$ is the corrected CMIP6 output.

The NASA Langley Fu–Liou radiative transfer model computes broadband solar shortwave and thermal longwave profiles of down-welling and up-welling flux accounting for gas absorption, and absorption and scattering by aerosols and clouds (Balmes et al 2021, Natarajan et al 2012). It is a highly modified version of the original model developed by Fu and Liou (Fu and Liou 1993). Atmospheric profiles (1000–0.1 hPa) of pressure (hPa), temperature (K), water vapor (g g^–1), and ozone (g g^–1) from MERRA2, surface albedo and emission from MODIS, and AOD and vertical profiles (1000–0.1 hPa) of black carbon (BC) and organic carbon (OC) from MERRA2 were served as the key input parameters. The Langley Fu–Liou radiative transfer model was adopted for rough evaluation of radiative effect caused by TCAL.

Based on the formation mechanism of the ATAL, the two primary factors causing the appearance of a TCAL are atmospheric dynamics (deep convection, upper-tropospheric anticyclone) and surface pollutants. Strong deep convection -controlled regions characterized by heavy precipitation are shown in figure 1(a), mainly over the land and ocean in the region 30° N–30° S, with South and East Asia including India and China having the most severe anthropogenic pollution (figures 1(b) and S2). In contrast, South America, Africa, and Southeast Asia are covered mainly by forest and grassland with pollution arising mainly from natural wildfires (figures 1(c) and (d)). There is relatively little atmospheric pollution over the Caribbean and other ocean areas. As shown in figure 1(b), the forest and grassland over South America and Africa would emit large amount of CAs through wildfire burning, with the considerable anthropogenic emission for CAs over Asia. The CAs, as one of aerosol components, have been adopted to identify features of ATAL (Yuan et al 2019a, Lau and Kim 2022). The presence of upper-tropospheric CAs, as indicated by multiple data from MERRA2 reanalysis, CMIP6 simulation, and OMPS/LP satellite retrieval (figures 2 and 3), indicates the occurrence of TCALs over Asia, South America, and Africa. Deep convections are strong in these areas, and surface pollution levels are high (figures 1(e)–(m) and 4).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The appearance of a TCAL follows a distinct seasonal pattern (figure 2). In Asia, the TCAL typically occurs during the strongest deep convection season, reaching its peak in July–August (figure 2(a)), as previously demonstrated (Yuan et al 2019a). However, during other seasons such as March–May, which are characterized by large biomass burning and weak deep convection, the TCAL phenomenon is absent over Asia, despite the presence of higher surface pollutant emissions compared to July–August (figures 2(d) and (g)). On the other hand, in South America, the TCAL is observed between October–March, when deep convection is strong and there is extensive surface emissions of CAs (figures 2(b), (e) and (h)). The intensity of the South American TCAL peaks during October–December due to variations in location and intensity of surface emission and deep convection (figures S3–S6). It is important to note that the TCAL does not occur during the weak deep convection season. The results in Africa are similar to those in South America. The African TCAL appears from November to March, which coincides with the deep convection season, rather than May–September (figures 2(c), (f) and (i)). During November–December, the African TCAL reaches its highest intensity. It should be mentioned that the peak intensities of the TCAL over South America and Africa do not align with the strongest deep convection season. In January–April, the region of accumulated CAs emissions over South America and Africa is not concentrated in the central area of strong deep convection (figures S4–S6). As a result, the deep convection fails to transport a substantial amount of CAs from the surface to the upper troposphere during this period. Therefore, this study will primarily focus on the seasons of July–August over Asia, October–December over South America, and November–December over Africa to explore the detailed characteristics of TCALs and the factors influencing their formation, despite these seasons not coinciding with the peak deep convection period (monsoon peak).

Spatial distributions and cross-sections of TCALs over Asia, South America, and Africa are shown in figure 3 for July–August, October–December, and November–December, respectively. Here, features of TCALs were examined mainly on the basis of MERRA2 and CMIP6 datasets, rather than the OMPS/LP dataset, owing to the low signal/noise ratio caused by cloud cover in the latter. Corresponding CAs AOD is shown in (figures 1(e)–(m)), and OLR, OMEGA, and the ∼200 hPa atmospheric circulation fields are shown in figure 4, indicating features of deep convention, vertical motion, and anticyclones (centers) in the upper troposphere.

Among these regions, South America exhibits the most intense surface CAs pollution, and Asia the least (figures 1(e)–(m)). There are two pollution centers in Asia: in India and China. During the strong deep convection period, upward vertical motion induced by deep convection carries surface pollutants into the upper troposphere and even the lower stratosphere (Yu et al 2017, Yuan et al 2019a). The highest TCAL intensity (concentration) was thus observed in South America, while intensity was weakest in Asia (based on MERRA2). TCAL concentrations in South America were 2–3 times those in Asia and Africa, consistent with patterns of surface pollution. Among the three regions, South America may not have the strongest deep convection activity, but it features the largest surface emission in comparison to the other two regions, which may play a dominant role in determining the intense intensity of TCAL. In order to quantitatively assess the impact of surface emission and deep convection on TCAL formation, further model simulations will be conducted. However, TCAL altitude was highest in Asia, centered at ∼120 hPa, followed by South America (∼200 hPa), and Africa (∼220 hPa). This is due to the Himalayas and the Tibetan Plateau causing the strongest deep convection in South Asia (Boos and Kuang 2010, Zhang et al 2012) with the upward vertical motion and upwelling velocities exceeding −0.06 Pa s⁻¹ in the core ascent zone. Deep convections in the other areas are relatively weak, even negligible in Africa relative to Asia. It follows that updrafts account for the suspended height of TCALs, and tropopause height alone does not determine the rising altitude of the TCAL (figures 4(g)–(i)).

Another interesting distinction between TCALs in different areas is their three-dimensional shape. The TCAL in Asia has the widest spatial coverage, spanning from the Iranian Plateau to eastern China through an extended upper-tropospheric anticyclonic circulation spanning >60° of longitude. This circulation is strengthened by the heating effect of the Himalayas and the Tibetan Plateau, as documented by Wu et al (2015). In contrast, anticyclonic circulation over South America and Africa is relatively weak, covering only ∼30° of longitude. This acts as a barrier, constraining the TCAL to a limited spread, mainly over central South America and Central Africa. Of the two TCAL centers in Asia, the stronger is over the Tibetan Plateau and the weaker is over southwest China, due to anthropogenic pollution (the primary emission source) being centered over small areas of dense population in India and western China and producing two narrow vertical transport conduits, consistent with Lau et al (2018). In comparison, surface pollutants in South America and Africa are derived mainly from widespread wildfires across forest and grassland, resulting in one broad vertical transport conduit and a single peak TCAL center. Overall, TCALs in Asia, South America, and Africa thus exhibit divergent spatial features owing to the combined effects of atmospheric dynamics and surface pollutant emissions.

The remarkable features of the TCALs result in divergent radiative effects. Based on MERRA2 reanalysis, we evaluated radiative effect during clear sky induced by CAs in the upper troposphere (TCAL) in Asia, South America, and Africa utilizing the Langley Fu–Liou radiative transfer model. As shown in figure 5(a), TCAL over Asia have warming effects on the climate system of +0.23 W m⁻² at the top of the atmosphere, whereas there are cooling effects of −0.54 and −0.20 W m⁻² over South America and Africa, respectively. These differences are related to CA composition, which includes BC and OC; the former has a high light-absorption capacity and the latter a predominantly light-scattering capacity. The relative CA contents of BC and OC determine RF magnitude and sign. The amount of BC in TCAL over Asia (AOD 0.0031) was higher than that in the South American (AOD 0.0017) and African (AOD 0.001) TCALs. The OC content (AOD 0.004) over Asia was roughly equivalent to the BC content, but it was much higher over South America (AOD 0.009) and Africa (AOD 0.004). Therefore, TCAL BC warming effects over South America and Africa were offset by OC cooling effects. In contrast to results at the top of the atmosphere, TCAL caused negative radiative effect at the surface for all three regions owing to absorption and scattering of sunlight by CA leading to less shortwave radiation reaching the surface. Among three regions, the surface radiative effect of TCAL was lowest in Africa. In the atmosphere, Asian TCAL had the strongest warming effect because of their higher BC contents. Consistent with the results of Gao et al (2023), our simulation indicated a positive Asian TCAL radiative effect at the top of the atmosphere rather than the negative values reported by Vernier et al (2015). We found that CA-induced warming effects were stronger than those caused by scattering and absorbing aerosols (total aerosol components; +0.15 W m⁻²; Gao et al 2023). These results thus reflect the importance of TCAL in climate systems relative to other types of aerosol. Moreover, we note that larger amounts of CAs will likely be released to the atmosphere by more frequent and intense wildfires caused by favorable meteorological conditions under global warming (Pu et al 2021, Xie et al 2022, Huang et al 2023, Zheng et al 2023). More attention should thus be paid to CAs or smoke from wildfires to evaluate their source, transport, radiation, and climate effects.

Zoom In Zoom Out Reset image size

The intensity of solar shortwave radiation varies with altitude, leading to diverse radiative effect of equivalent CA concentrations at different altitudes (figure S7). The aerosol usually distributed in the lower troposphere, this study however found it can also be effectively transported to the upper troposphere. To highlight the importance of TCAL in the upper troposphere and emphasize the differences in radiative effects of TCAL at different altitudes, we then examined equivalent CAs at three altitude levels: upper troposphere (300–100 hPa), middle troposphere (600–400 hPa), and lower troposphere (900–700 hPa) (figure 5(b)). In Asia, the equivalent CA levels changed from warming to cooling effects with decreasing altitude, indicating the relative importance of competitive scattering and absorption of radiation by BC and OC at different altitudes, which was consistent with the conclusions from Ban-Weiss et al (2012). However, in South America and Africa, equivalent CAs at different altitudes exhibited consistently negative radiative effects, but with different intensities. These changes from warming to cooling, and radiative effect intensity, result from competition between BC absorption and OC scattering (figure S7). However, when only one type of aerosol (BC or OC) is considered, radiative effect increases with altitude because of the stronger solar shortwave radiation at higher altitudes, thus highlighting the importance of TCALs to the climate system.