Southern West Africa (SWA) is an African pollution hotspot but a relatively poorly sampled region of the world. We present an overview of in situ aerosol optical measurements collected over SWA in June and July 2016 as part as of the DACCIWA (Dynamics-Aerosol-Chemistry-Clouds Interactions in West Africa) airborne campaign. The aircraft sampled a wide range of air masses, including anthropogenic pollution plumes emitted from the coastal cities, long-range transported biomass burning plumes from central and southern Africa and dust plumes from the Sahara and Sahel region, as well as mixtures of these plumes. The specific objective of this work is to characterize the regional variability of the vertical distribution of aerosol particles and their spectral optical properties (single scattering albedo: SSA, asymmetry parameter, extinction mass efficiency, scattering Ångström exponent and absorption Ångström exponent: AAE). The first findings indicate that aerosol optical properties in the planetary boundary layer were dominated by a widespread and persistent biomass burning loading from the Southern Hemisphere. Despite a strong increase in aerosol number concentration in air masses downwind of urban conglomerations, spectral SSA were comparable to the background and showed signatures of the absorption characteristics of biomass burning aerosols. In the free troposphere, moderately to strongly absorbing aerosol layers, dominated by either dust or biomass burning particles, occurred occasionally. In aerosol layers dominated by mineral dust particles, SSA varied from 0.81 to 0.92 at 550 nm depending on the variable proportion of anthropogenic pollution particles externally mixed with the dust. For the layers dominated by biomass burning particles, aerosol particles were significantly more light absorbing than those previously measured in other areas (e.g. Amazonia, North America), with SSA ranging from 0.71 to 0.77 at 550 nm. The variability of SSA was mainly controlled by variations in aerosol composition rather than in aerosol size distribution. Correspondingly, values of AAE ranged from 0.9 to 1.1, suggesting that lens-coated black carbon particles were the dominant absorber in the visible range for these biomass burning aerosols. Comparison with the literature shows a consistent picture of increasing absorption enhancement of biomass burning aerosol from emission to remote location and underscores that the evolution of SSA occurred a long time after emission.
The results presented here build a fundamental basis of knowledge about the aerosol optical properties observed over SWA during the monsoon season and can be used in climate modelling studies and satellite retrievals. In particular and regarding the very high absorbing properties of biomass burning aerosols over SWA, our findings suggest that considering the effect of internal mixing on absorption properties of black carbon particles in climate models should help better assess the direct and semi-direct radiative effects of biomass burning particles.
Atmospheric aerosols play a crucial role in the climate system by altering the radiation budget through scattering and absorption of solar radiation and by modifying cloud properties and lifetime. Yet considerable uncertainties remain about the contribution of both natural and anthropogenic aerosol to the overall radiative effect (Boucher et al., 2013). Large uncertainties are related to the complex and variable properties of aerosol particles that depend on the aerosol source and nature as well as on spatial and temporal variations. During transport in the atmosphere, aerosol particles may undergo physical and chemical aging processes altering the composition and size distribution and henceforth the optical properties and radiative effects. The capability of reproducing this variability in climate models represents a real challenge (Myhre et al., 2013; Stier et al., 2013; Mann et al., 2014). Therefore, intensive experimental observations in both aerosol source and remote areas are of paramount importance for constraining and evaluating climate models.
Key parameters from a climate perspective are the aerosol vertical
distribution and respective spectral optical properties. Radiative transfer
codes commonly incorporated in climate models and in satellite data
retrieval algorithms use single scattering albedo (SSA), mass extinction
efficiency (MEE) and asymmetry factor (
Southern West Africa (SWA) is one of the most climate-vulnerable regions in
the world, where the surface temperature is expected to increase by
In West Africa, most of the aerosol–radiation interaction studies focused on optical properties of dust and biomass burning aerosols in remote regions far from major sources of anthropogenic pollution aerosol. They include ground-based and airborne field campaigns such as DABEX (Dust and Biomass Experiment; Haywood et al., 2008), AMMA (Analysis Multidisciplinary of African Monsoon; Lebel et al., 2010), DODO (Dust Outflow and Deposition to the Ocean; McConnell et al., 2008), SAMUM-1 and SAMUM-2 (Saharan Mineral Dust Experiment; Heintzenberg, 2009; Ansmann et al., 2011) and AER-D (AERosol Properties – Dust; Ryder et al., 2018). These projects concluded that the influence of both mineral dust and biomass burning aerosols on the radiation budget is significant over West Africa, implying that meteorological forecast and regional/global climate models should include their different radiative effects for accurate forecasts and climate simulations. Over the Sahel region, Solmon et al. (2008) have highlighted the high sensitivity of mineral dust optical properties to precipitation changes at a climatic scale. However, the optical properties of aerosol particles in the complex chemical environment of SWA are barely studied. This is partly due to the historically low level of industrial developments of the region. Motivated by the quickly growing cities along the Guinean coast, the study of transport, mixing, and feedback processes of aerosol particles is therefore very important for better quantification of aerosol radiative impact at the regional scale and improvement of climate and numerical weather prediction models.
In this context, the DACCIWA (Dynamics-Aerosol-Chemistry-Clouds Interactions in West Africa; Knippertz et al., 2015b) campaign, designed to characterize both natural and anthropogenic emissions over SWA, provides important and unique observations of aerosols in a region much more affected by anthropogenic emissions than previously thought. A comprehensive field campaign took place in June–July 2016 including extensive ground-based (Kalthoff et al., 2018) and airborne measurements (Flamant et al., 2018b). In this study, we present an overview of in situ airborne measurements of the vertical distribution of aerosol particles and their spectral optical properties acquired with the ATR-42 French research aircraft over the Guinean coast.
Section 2 presents the flight patterns, instrumentation and data analysis. Section 3 provides an overview of the aerosol microphysical and optical properties. The impact of aging and mixing processes on aerosol optical properties is discussed in Sect. 4 before conclusions are presented in Sect. 5.
This analysis focuses on flight missions conducted by the ATR-42 aircraft of SAFIRE (Service des Avions Français Instrumentés pour la Recherche en Environnement – the French aircraft service for environmental research) over the Gulf of Guinea and inland. A full description of flight patterns during DACCIWA is given in Flamant et al. (2018b). Here we present results from 15 flights focused on the characterization of anthropogenic pollution, dust and biomass burning plumes. The flight tracks are shown in Fig. 1 and a summary of flight information is provided in Table A1 in the Appendix. The sampling strategy generally consisted of two parts: first, vertical soundings were performed from 60 m up to 8 km above mean sea level (a.m.s.l.) to observe and identify interesting aerosol layers. Subsequently, the identified aerosol layers were probed with the in situ instruments by straight levelled runs (SLRs) at fixed flight altitudes.
Tracks of the 15 flights analysed in this study. The colours indicate aircraft flight sampling layers dominated by biomass burning (green), mineral dust (orange), mixed dust–biomass burning (red), anthropogenic pollution (blue) and background particles (grey) from both vertical profiles (squares) and straight and level runs (SLRs; dots).
The ATR-42 aircraft was equipped with a wide variety of instrumentation
performing gas and aerosol measurements. The measured meteorological
parameters include temperature, dew point temperature, pressure, turbulence,
relative humidity, as well as wind speed and direction. Gas-phase species
were sampled through a rear-facing
The total number concentration of particles larger than 10 nm (
The particle extinction coefficient (
In order to determine the history of air masses prior to aircraft sampling,
backward trajectories and satellite images were used. The trajectories were
computed using the Hybrid Single Particle Lagrangian Integrated Trajectory
Model (HYSPLIT) and the National Centers for Environmental Prediction (NCEP)
Global Data Assimilation System (GDAS) data with 0.5
In the following, extensive aerosol parameters (concentrations, scattering,
absorption and extinction coefficients) are converted to standard
temperature and pressure (STP) using
Figure A1 and Table 1 show the iterative procedure and the equations used to calculate the aerosol microphysical and optical parameters as briefly explained below.
The particle number concentration in the coarse mode (
Aerosol microphysical and optical properties derived in this work.
For optical calculations, the 3 SAE depends on the size of the particles. Generally, it is lower than 0 for
aerosols dominated by coarse particles, such as dust aerosols, but it is
higher than 0 for fine particles, such as anthropogenic pollution or biomass
burning aerosol (Seinfeld and Pandis, 2006; Schuster et al., 2006). AAE provides information about the chemical composition of atmospheric
aerosols. BC absorbs radiation across the whole solar spectrum with the same
efficiency; thus, it is characterized by AAE values around 1. Conversely, mineral
dust particles show strong light absorption in the blue to ultraviolet
spectrum, leading to AAE values up to 3 (Kirchstetter et al., 2004; Petzold et al., 2009). SSA describes the relative importance of scattering and absorption for
radiation. Thus, it indicates the potential of aerosols to cool or
warm the lower troposphere. MEE represents the total light extinction per unit mass concentration of
aerosol. The estimates of MEE assume mass densities of 2.65 g cm
Data were screened in order to isolate plumes dominated by anthropogenic
pollution from urban emissions, biomass burning and mineral dust particles,
resulting in a total number of 19, 12 and 8 genuine plume interceptions,
respectively, across the 15 flights. As shown in Fig. 2, identification of
the plumes was based on a combination of CO and
Absorption Ångström exponent (AAE) as a function of the
ratio
Backward trajectories for the analysed aerosol layers.
Trajectories date back 10 d for panel
The guidelines for classification are as follows.
Figure 4 shows a statistical analysis of
Vertical layering of aerosols and meteorological variables for
profiles for which aerosols dominated by biomass burning (green), dust
(orange), mixed dust–biomass burning (red), anthropogenic pollution (blue)
and background particles (black) were detected. The panels show profiles of
The observed wind profiles highlight the presence of several distinct layers
in the lower troposphere. For cases related to dust, urban pollution and
background condition, we clearly observe the monsoon layer up to 1.5 km a.m.s.l.,
which is characterized by weak to moderate wind speeds (2 to 10 m s
The vertical distribution of aerosol particles was very inhomogeneous, both
across separate research flights and between individual plumes encountered
during different periods of the same flight. Measurements of aerosols within
this analysis cover a broad geographic region, as shown in Fig. 1, which
may explain some of the variability. SWA is subject to numerous
anthropogenic emission sources (e.g. road traffic, heavy industries, open
agriculture fires) coupled to biogenic emissions from the ocean and
forests. These resulting large emissions are reflected in the high
variability of
A prominent feature in the vertical profiles is the presence of fine
particles up to 2.5 km a.m.s.l. outside of biomass burning or dust events.
Figure 5a shows the range of variability of the number and volume size distributions measured during DACCIWA. These are extracted from the SLRs identified in Fig. 1. Figure 5b shows the same composite distribution normalized by CO concentration in order to account for differences in the amount of emissions from combustion sources.
Considerable variability in the number concentration of the size distributions, up to approximately 2 orders of magnitude, was observed for a large fraction of the measured size range. The size distributions varied both for different aerosol types and for a given aerosol class. This reflects the relatively wide range of different conditions that were observed over the region, in terms of sources, aerosol loading, and lifetimes of plumes.
In particular for ultrafine particles with diameters below 100 nm, large
differences were observed, with an increase as large as a factor of 50 in
urban plumes, which reflects concentration increase from freshly formed
particles. Interestingly elevated number concentrations of these
small-diameter particles were also observed in some dust layers. Comparing
the particle size distribution of the different dust plumes sampled during
the field campaign, a variation as large as a factor of 20 in the number
concentration of ultrafine particles is found (i.e. Fig. 4). Their
contribution decreased with height, as reflected by the higher small particle
number recorded in dust plumes below 2.5 km a.m.s.l. (Figs. 4b and 5b). As the
composite urban size distributions showed a relatively similar ultrafine
mode centered at 50 nm, dust layers have most likely significant
contributions from anthropogenic pollution aerosol freshly emitted in SWA.
The ultrafine mode was not observed in biomass burning size distributions,
even though dust and biomass burning plumes were sampled in the same
altitude range. We interpret this observation with dust plumes transported
below 2.5 km a.m.s.l. that were sampled over the region of Savè
(8
The accumulation mode was dominated by two modes centered at
The number concentration of large super-micron particles was strongly
enhanced in the mineral dust layers. The peak number concentration displayed
a broad shape at
SSA is one of the most relevant intensive optical properties because it describes the relative strength of the aerosol scattering and absorption capacity and is a key input parameter in climate models (Solmon et al., 2008). Figure 6 shows the spectral SSA for the different SLRs considered in this study.
The highest absorption (lowest SSA) at all three wavelengths was observed for biomass burning aerosols. SSA values ranged from 0.69 to 0.78 at 440 nm, 0.71 to 0.77 at 550 nm and 0.65 to 0.76 at 660 nm. This is on the low side of the range of values (0.73–0.93 at 550 nm) reported over West Africa during DABEX for biomass burning plumes mixed with variable proportion of mineral dust (Johnson et al., 2008). No clear tendency was found for the spectral dependences of SSA, which in some of the cases decreased with wavelength and in others were very similar to each other at all three wavelengths.
SSA values of anthropogenic pollution aerosols were generally intermediate in magnitude, with median values of 0.81 at 440 nm, 0.82 at 550 nm and 0.82 at 660 nm. Our data show that the value of SSA varied significantly for the different plumes. Some pollution aerosols absorb almost as strongly as biomass burning aerosols with SSA(550 nm) values as low as 0.72, whereas the highest SSA(550 nm) value observed was 0.86. In addition, the absorption properties of urban aerosol varied greatly between the sampled plumes for smoke of an apparently identical geographic origin. For example, we measured SSA(550 nm) values from 0.72 to 0.82 in the Accra pollution outflow. The variability in SSA values may be due to the possible contribution of emissions from different cities to the sampled pollution plumes (Deroubaix et al., 2019), thus having different combustion sources and chemical ages. The flat spectral dependence of SSA appears to be anomalous for anthropogenic pollution aerosols, as SSA has been shown to decrease with increasing wavelength for a range of different urban pollution plumes (Dubovick et al., 2002; Petzold et al., 2011; Di Biagio et al., 2016; Shin et al., 2019).
The magnitude of SSA increased at the three wavelengths when dust events
occurred. Large variations in SSA were obtained, with values ranging from
0.76 to 0.92 at 440 nm, from 0.81 to 0.94 at 550 nm and from 0.81 to 0.97 at 660 nm. The
measurement of SSA is highly dependent on the extent to which the coarse mode
is measured behind the aerosol sampling inlet. Denjean et al. (2016) found that the absolute
error associated with SSA,
Single scattering albedo, mass extinction efficiency (in
m
As shown in Table 2, the observed variability of SSA reflects a large
variability for MEE at 550 nm, which spans a wide range from 0.38 to 1.37, 1.45 to 1.92 and 1.24 to 4.83 m
This analysis includes sampled aerosols originating from different source regions and having undergone different aging and mixing processes, which could explain some of the variability. The impact of these factors on the magnitude and spectral dependence of optical parameters will be investigated in the following section.
Figure 7 shows the vertical distribution of SSA, SAE and the
Vertical distribution of
One of the critical factors in the calculation of aerosol direct and
semi-direct radiative effects is the mixing state of the aerosols, which can
significantly affect absorbing properties. There were no direct
observational constraints available on this property during the DACCIWA
airborne campaign. However, we investigated the probable aerosol mixing
state by calculating composite SSA from the aerosol size distribution. On
the basis of Fig. 5, dust size distributions showed only minor
discrepancies in the mean and standard deviation of the coarse mode but
significant differences in the balance between the fine and coarse modes, which
suggests low internal mixing of dust with other atmospheric species. The
size distributions of mixed dust pollution have been deconvoluted by
weighting the size distributions of mineral dust and anthropogenic pollution
aerosol averaged over the respective flights. This assumes that dust was
externally mixed with the anthropogenic pollution particles and assumes a
homogeneous size distribution for the dust and anthropogenic pollution
aerosol throughout a flight.
SSA, SAE and
In the boundary layer, the similar SSA and SAE in anthropogenic pollution and
background plumes suggest that background aerosol may be rather called
background pollution originating from a regional background source in the
far field. Our analysis of the spectral dependence of SSA showed no apparent
signature of anthropogenic pollution aerosols (see Sect. 3.3) despite a
strong increase in aerosol number concentrations in air masses crossing
urban centres (see Sect. 3.2). This can be explained by two factors.
First, the majority of accumulation-mode particles were present in the
background, while the large proportion of aerosols emitted from cities
resided in the ultrafine-mode particles that have fewer scattering
efficiencies (Fig. 5). Second, large amounts of absorbing aerosols in the
background can minimize the impact of further increase in absorbing
particles to the aerosol load. The high CO values (
The optical properties of aerosols are determined by either the aerosol chemical composition, the aerosol size distribution, or both. Changes in the size distribution of biomass burning aerosol due to coagulation and condensation have been shown to alter the SSA, as particles increase towards sizes for which scattering is more efficient (Laing et al., 2016). Variations in particle chemical composition, caused by source emissions and aging processes associated with gas-to-particle transformation and internal mixing, have been shown to change the SSA (Abel et al., 2003; Petzold et al., 2011).
In order to determine the contributions from size distribution and chemical
composition to the variation of SSA in biomass burning plumes, SSA is presented as a
function of SAE and
Contribution to single scattering albedo
In contrast, Fig. 8b shows that there was a consistent decrease in SSA with
increasing
Compared with past in situ measurements of aged biomass burning aerosol,
SSA values over SWA (0.71–0.77 at 550 nm) are at the lower end of those
reported worldwide (0.73–0.99 at 550 nm) (Magi et al., 2003; Reid et al., 2005; Johnson et al., 2008; Corr et al., 2012; Laing et al., 2016). This can be attributed in part
to the high flaming versus smouldering conditions of African smoke producing
more BC particles (Andreae and Merlet, 2001; Reid et al., 2005), which inherently have low SSA compared to other regions
(Dubovick et al., 2002). However, SSA values over SWA are significantly lower than the range
reported near emission sources in sub-Saharan Africa and over the south-eastern
Atlantic, where values span over 0.84–0.90 at 550 nm (Haywood et al., 2003b; Pistone et al., 2019). Recent observations
carried out on Ascension Island to the south-west of the DACCIWA region
showed that smoke transported from central and southern African fires can be
very light absorbing over the July–November burning season, but SSA values were
still higher (
Currently there are few field measurements of well-aged biomass burning emissions. Our knowledge of biomass burning aerosol primarily comes from laboratory experiments and near-field measurements taken within a few hours of a wildfire (Abel et al., 2003; Yokelson et al., 2009; Adler et al., 2011; Haywood et al., 2003b; Vakkari et al., 2014; Zhong and Jang, 2014; Forrister et al., 2015; Laing et al., 2016; Zuidema et al., 2018). With the exception of the study by Zuidema et al. (2018) over the south-eastern Atlantic, it is generally found that the aged biomass burning aerosol particles are less absorbing than freshly emitted aerosols due to a combination of condensation of secondary organic species and an additional increase in size by coagulation. This is in contrast to our results showing that SSA of biomass burning aerosols was significantly lower than directly after emission and that the evolution of SSA occurred a long time after emission.
There are three possible explanations for these results. First, one must consider sample bias. As regional smoke ages, it can be enriched by smoke from other fires that can smoulder for days, producing large quantities of non-absorbing particles, thereby increasing the mean SSA (Reid et al., 2005; Laing et al., 2016). However, during DACCIWA, biomass burning plumes were transported over the Atlantic Ocean and were probably less influenced by multiple fire emissions. Second, there is evidence that fresh BC particles become coated with sulfate and organic species as the plume ages in a manner that enhances their light absorption (Lack et al., 2012; Schwarz et al., 2008). Finally, organic particles produced during the combustion phase can be lost during the transport through photobleaching, volatilization and/or cloud-phase reactions (Clarke et al., 2007; Lewis et al., 2008; Forrister et al., 2015), which is consistent with the low SSA and AAE values we observed. Assessing whether these aging processes impact the chemical components and henceforth optical properties of transported biomass burning aerosol would need extensive investigation of aerosol chemical composition that will be carried out in a subsequent paper.
This paper provides an overview of in situ airborne measurements of vertically
resolved aerosol optical properties carried out over SWA during the DACCIWA
field campaign in June–July 2016. The peculiar dynamics of the region lead
to a chemically complex situation, which enabled sampling of various air
masses, including long-range transport of biomass burning from central
Africa and dust from Sahelian and Saharan sources, local anthropogenic
plumes from the major coastal cities, and mixtures of these different
plumes. This work fills a research gap by providing, firstly, key climate-relevant aerosol properties (SSA, MEE,
The aerosol vertical structure was very variable and mostly influenced by
the origin of air mass trajectories. While aerosol extinction coefficients
generally decreased with height, there were distinct patterns of profiles
during dust and biomass burning transport to SWA. When present, enhanced
values of extinction coefficients up to 240 Mm
The aerosol light absorption in dust plumes was strongly enhanced as a result of this mixing. We find a decrease in SSA(550 nm) from 0.92 to 0.81 for dust affected by anthropogenic pollution mixing compared to the situation in which the dust plumes moved at higher altitudes across SWA. Comparison of the particle size distributions of the different dust plumes showed a large contribution of externally mixed fine-mode particles in mixed layers, while there was no evidence of internal mixing of coarse particles. Concurrent optical calculations by deconvoluting size-distribution measurements in mixed layers and assuming an external mixing state allowed us to reproduce the observed SSAs. This implies that an external mixing would be a reasonable assumption to compute aerosol direct and semi-direct radiative effects in mixed dust layers.
Despite a strong increase in aerosol number concentration in air masses crossing urban conglomerations, the magnitude of the spectral SSAs was comparable to the background. Enhancements of light absorption properties were seen in some pollution plumes but were not statistically significant. A persistent spectral signature of biomass burning aerosols in both background and pollution plumes highlights that the aerosol optical properties in the boundary layer were strongly affected by the ubiquitous biomass burning aerosols transported from central Africa (Menut et al., 2018; Haslett et al., 2019). The large proportion of aerosols emitted from the cities of Lomé, Accra and Abidjan that resided in the ultrafine-mode particles have limited impact on already elevated amounts of accumulation-mode particles having a maximal absorption efficiency. As a result, in the boundary layer, the contributions from local city emissions to aerosol optical properties were of secondary importance at regional scale compared with this large absorbing aerosol mass. While local anthropogenic emissions are expected to rise as SWA is currently experiencing major economic and population growth, there is increasing evidence that climate change is increasing the frequency and distribution of fire events (Jolly et al., 2015). In terms of future climate scenarios and accompanying aerosol radiative forcing, whether the large biomass burning events that occur during the monsoon season would limit the radiative impact of increasing anthropogenic emissions remains an open and important question.
The SSA values of biomass burning aerosols transported in the free troposphere
were very low (0.71–0.77 at 550 nm) and have only rarely been observed in
the atmosphere. The variability in SSA was mainly controlled by the variability
in aerosol composition (assessed via
We believe the set of DACCIWA observations presented here is representative of the regional mean and variability in aerosol optical properties that can be observed during the monsoon season over SWA, as the main dynamical features were in line with the climatology (Knippertz et al., 2017). This is why results from the present study will serve as input and constraints for climate modelling to better understand the impact of aerosol particles on the radiative balance and cloud properties over this region and will also substantially support remote sensing retrievals.
Summary of flight information. All flights were conducted during 2016.
Data inversion procedure to calculate the aerosol microphysical and optical parameters.
All data used in this study are publicly available at the AERIS Data and
Service Center, which can be found at
CD conducted the analysis of the data and wrote the paper. CD, TB, FB, NM, AC, PD, JB, RD, KS and AS operated aircraft instruments and processed and/or quality-controlled data. MM provided expertise on aerosol–climate interaction processes. CF and PK were the PIs, who led the funding application and coordinated the DACCIWA field campaign. All the co-authors contributed to the writing of the paper.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Results of the project “Dynamics-aerosol-chemistry-cloud interactions in West Africa” (DACCIWA) (ACP/AMT inter-journal SI)”. It is not associated with a conference.
The research leading to these results has received funding from the European
Union 7th Framework Programme (FP7/2007–2013) under grant agreement no. 603502 (EU project DACCIWA: Dynamics-Aerosol-Chemistry-Cloud Interactions in
West Africa). The European Facility for Airborne Research (EUFAR,
This research has been supported by the FP7 Environment (DACCIWA (grant no. 603502)).
This paper was edited by Andreas Petzold and reviewed by two anonymous referees.