Volume 12, Issue 1 p. 13-18
Special Issue Article
Open Access

Atmospheric composition of West Africa: highlights from the AMMA international program

Céline H. Mari

Corresponding Author

Céline H. Mari

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Laboratoire Aerologie, Universite de Toulouse, CNRS, 31400 Toulouse, France.Search for more papers by this author
Claire E. Reeves

Claire E. Reeves

School of the Environmental Sciences, University of East Anglia, Norwich, UK

Search for more papers by this author
Katherine S. Law

Katherine S. Law

UPMC, Université de Paris 06, Université Versailles St-Quentin, CNRS/INSU, LATMOS-IPSL, Paris, France

Search for more papers by this author
Gérard Ancellet

Gérard Ancellet

UPMC, Université de Paris 06, Université Versailles St-Quentin, CNRS/INSU, LATMOS-IPSL, Paris, France

Search for more papers by this author
Maria Dolores Andrés-Hernández

Maria Dolores Andrés-Hernández

Institute of Environmental Physics, University of Bremen, Bremen, Germany

Search for more papers by this author
Brice Barret

Brice Barret

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Search for more papers by this author
Joëlle Bechara

Joëlle Bechara

Laboratoire Interuniversitaire des Systèmes Atmosphèriques, Universités de Paris 12 et Paris 7, CNRS, Créteil, France

Search for more papers by this author
Agnès Borbon

Agnès Borbon

Laboratoire Interuniversitaire des Systèmes Atmosphèriques, Universités de Paris 12 et Paris 7, CNRS, Créteil, France

Search for more papers by this author
Idir Bouarar

Idir Bouarar

UPMC, Université de Paris 06, Université Versailles St-Quentin, CNRS/INSU, LATMOS-IPSL, Paris, France

Search for more papers by this author
Francesco Cairo

Francesco Cairo

ISAC, Institute for Atmospheric Sciences and Climate, National Research Council, Italy

Search for more papers by this author
Roisin Commane

Roisin Commane

School of Chemistry, University of Leeds, Leeds, UK

Search for more papers by this author
Claire Delon

Claire Delon

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Search for more papers by this author
Matthew J. Evans

Matthew J. Evans

School of the Environment, University of Leeds, Leeds, UK

Search for more papers by this author
Federico Fierli

Federico Fierli

ISAC, Institute for Atmospheric Sciences and Climate, National Research Council, Italy

Search for more papers by this author
Cédric Floquet

Cédric Floquet

School of Chemistry, University of Leeds, Leeds, UK

Search for more papers by this author
Corinne Galy-Lacaux

Corinne Galy-Lacaux

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Search for more papers by this author
Dwayne E. Heard

Dwayne E. Heard

School of Chemistry, University of Leeds, Leeds, UK

Search for more papers by this author
Carine D. Homan

Carine D. Homan

Department of Physics, University of Wuppertal, Germany

Search for more papers by this author
Trevor Ingham

Trevor Ingham

School of Chemistry, University of Leeds, Leeds, UK

Search for more papers by this author
Niels Larsen

Niels Larsen

Danish Meteorological Institute, Danish Climate Center, Copenhagen, Denmark

Search for more papers by this author
Alastair C. Lewis

Alastair C. Lewis

National Centre for Atmospheric Science, University of York, Heslington, York, UK

Search for more papers by this author
Catherine Liousse

Catherine Liousse

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Search for more papers by this author
Jennifer G. Murphy

Jennifer G. Murphy

Department of Chemistry, University of Toronto, Ontario, Canada

Search for more papers by this author
Emiliano Orlandi

Emiliano Orlandi

ISAC, Institute for Atmospheric Sciences and Climate, National Research Council, Italy

Search for more papers by this author
David E. Oram

David E. Oram

School of the Environmental Sciences, University of East Anglia, Norwich, UK

Search for more papers by this author
Marielle Saunois

Marielle Saunois

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

National Center for Atmospheric Research, Boulder, Colorado, USA

Search for more papers by this author
Dominique Serça

Dominique Serça

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Search for more papers by this author
David J. Stewart

David J. Stewart

Department of Chemistry, University of Reading, Reading, UK

Search for more papers by this author
Daniel Stone

Daniel Stone

School of Chemistry, University of Leeds, Leeds, UK

School of the Environment, University of Leeds, Leeds, UK

Search for more papers by this author
Valérie Thouret

Valérie Thouret

Laboratoire Aerologie, University of Toulouse, CNRS, Toulouse, France

Search for more papers by this author
Peter van Velthoven

Peter van Velthoven

Royal Netherlands Meteorological Institute, De Bilt, The Netherlands

Search for more papers by this author
Jason E. Williams

Jason E. Williams

Royal Netherlands Meteorological Institute, De Bilt, The Netherlands

Search for more papers by this author
First published: 16 August 2010
Citations: 19

Abstract

The atmospheric composition of West Africa reflects the interaction of various dynamical and chemical systems (i.e. biogenic, urban, convective and long-range transport) with signatures from local to continental scales. Recent measurements performed during the African Monsoon Multidisciplinary Analyses (AMMA) observational periods in 2005 and 2006 provide new data which has allowed new insight into the processes within these systems that control the distribution of ozone and its precursors. Using these new data and recently published results, we provide an overview of these systems with a particular emphasis on ozone distributions over West Africa during the wet season. Copyright © 2010 Royal Meteorological Society

1. Background

The West African subregion is an important provider of ozone and aerosols, which are radiatively active components in the climate system. Prior to the African Monsoon Multidisciplinary Analyses (AMMA) program, our knowledge about the distribution of O3 and its precursors over West Africa was limited. Measurements performed during the AMMA observational periods in 2005 and 2006 revealed a variety of new dynamical and chemical mechanisms that control the distribution of ozone and its precursors in this subregion. Details of the field campaigns are available in Lebel et al. (2010).

During the dry season (boreal winter), West Africa is marked by strong emissions of pollutants from biomass burning. During the wet season (typically from May to September), the region is influenced by mesoscale convective systems (MCSs), which impact the composition of the atmosphere through several processes (i.e. rapid vertical transport of gases and aerosols to the upper troposphere (UT), heterogeneous processing, emissions of NOx by lightning and alteration of the land surface wetness affecting the liberation of NOx from soils). Combustion of fuelwood for domestic energy is a continual source of air pollution primarily in urban areas. Vegetated regions emit large amounts of biogenic organic compounds which influence the production of ozone. During the AMMA program, an experimental strategy was set up to quantify these processes and to understand their impact at the global scale.

2. Tropical biogenically dominated environments

In the lower troposphere over West Africa during the wet season, the ozone distribution shows a significant south–north gradient with lower values over forested regions and higher values north of 12°N (Adon et al., 2010; Reeves et al., 2010) (Figure 1). From dense rain forest in the coastal belt to the sub-Sahelian savanna in the north, there are 72 million hectares of forest in West Africa. Dry deposition loss of ozone to vegetation is the main driver of the ozone minimum over the forested areas, but factors relating to the biogenic emissions also affect the observed ozone latitudinal profile (Saunois et al., 2009).

Details are in the caption following the image

(Left) Observed isoprene-mixing ratio (in pptv) along the BAe-146 aircraft tracks below 700 m (refer Murphy et al., 2010 for details on aircraft measurements). (Right) Latitude–altitude distribution of ozone simulated by a mesoscale model (bottom) and observed by the BAe-146 aircraft (top) in August 2006. The green bar marks the latitudinal extension of the vegetated area. The red arrow indicates the advection to northern latitudes by the nocturnal boundary layer jet (adapted from Saunois et al., 2009)

Vegetated regions, south of 10°N, emit large amounts of biogenic volatile organic compounds (VOCs) (Ferreira et al., 2010; Murphy et al., 2010) (Figure 1). Emissions from vegetation are dependent both on environmental conditions and plant type; high emission rates of isoprene were observed for West African native plants, while a non-native plant originating from South America was found to be more important for monoterpenes (Saxton et al., 2007). Aircraft measurements of OH and HO2 radicals (Commane et al., 2010) indicate that the maximum concentrations of both species occur over the forested region. HO2 is controlled by relatively simple photochemical processes (Stone et al., 2010). In contrast, the high reactivity of the short-lived biogenic VOCs leads to model underestimates of OH, similar to that found in other low NOx regions of the world impacted by biogenic VOCs (Saunois et al., 2009). Conversely, longer-lived and secondary organics can be oxidized north of the forested area where they are transported and/or produced and contribute to the ozone maximum there (Saunois et al., 2009) (Figure 1).

Enhanced concentrations of NOx and O3 observed in the boundary layer over semiarid Sahelian regions following the passage of precipitating MCSs (Stewart et al., 2008) were interpreted as being due to the release of NOx from soils. A mesoscale model study in which NOx emissions from soils were defined by an algorithm driven by various environmental parameters (e.g. soil moisture and pH) supported these conclusions (Delon et al., 2008). As part of the IDAF network, long-term measurements over diverse African ecosystems confirmed that, in the wet season, concentrations of nitrogen components were highest over the dry savannas (Adon et al., 2010) and that the nitrogen cycle of the whole Sahelian region is impacted by these strong NOx emissions together with the ammonia source from animals (Galy-Lacaux et al., 2009). Saunois et al. (2009) showed that the soil NOx emissions combined with northward advection of biogenic VOCs play a key role in producing enhanced ozone concentrations over the dry savanna regions, with production rates of up to about 7 ppbv/day.

This large-scale impact of biogenic emissions was also verified by Williams et al. (2009) who found that biogenic VOCs released from Africa are estimated to contribute 2–4% of the global burden of VOC and that 2–45% of tropospheric O3 over equatorial Africa may come from African soil NOx emissions. Further, these emissions also contribute to enhanced ozone production over the tropical Atlantic downwind of West Africa (Williams et al., 2009).

3. Urban environments and air quality

Air pollution, particularly in urban centres, is an emerging issue for human health in many West African countries. Increasing levels of toxic pollutants are a result of industrial emissions and vehicle exhausts as well as the burning of coal, wood or other fuels to meet domestic energy requirements. Exceptionally high O3 concentrations (up to 284 ppbv at 1 km altitude) observed during the dry season by the ozone sounding network in Cotonou, Benin, have been linked to an unusual combination of sources including biomass burning, urban pollution and the petrochemical industry (Minga et al., 2010). Hopkins et al. (2009) reported top–down emissions estimate for the Lagos megacity in Nigeria based on aircraft measurements. Annual emission fluxes for NOx were found to be comparable with previous bottom–up estimates for other developing megacities, whereas VOC and CO emissions per capita were among the highest ever reported. Interestingly, measured O3 levels were not significantly elevated in this case, possibly due to titration in polluted conditions.

4. Convective environments

MOZAIC commercial aircraft data over the West African continent had previously shown that convection uplifts O3-poor air into the UT and contributes to an observed O3 minimum at 12–14 km (Sauvage et al., 2007; Saunois et al., 2008). This was further confirmed during the AMMA by ozonesondes (Cairo et al., 2010; Thouret et al., 2009) (Figure 2) and aircraft observations (Ancellet et al., 2009).

Details are in the caption following the image

Vertical ozone-mixing ratio profiles from 17 ozone soundings in Niamey (Niger) between 27 July and 25 August 2006 (courtesy N. Larsen, adapted from Cairo et al., 2009)

Using aircraft measurements, Bechara et al. (2009) observed up to three times higher concentrations of VOCs in the UT during convective conditions compared to non-convective conditions and model studies suggest that the UT is frequently perturbed by MCSs up to an altitude of about 14 km (Law et al., 2010). In addition, the MCS can produce NOx from lightning, but NOx production per standard stroke over West Africa was found to be 40% lesser than the thunderstorms over northern Australia and southern Germany (Höller et al., 2009). Despite this, NO concentrations were found to be enhanced in the UT (Figure 3).

Details are in the caption following the image

Vertical profiles of NO-mixing ratios observed during the AMMA Special Observational Period in August 2006 by the DLR Falcon 20 (red) and simulated by four chemistry transport models: MOCAGE, LMDz-INCA, TM4 and TOMCAT for (Left) observations that have been impacted by MCSs in the previous 3–4 days and (right) observations with no recent MCSs impact (courtesy I. Bouarar; adapted from Barret et al., 2010)

Moreover, Andrés-Hernández et al. (2009), on occasions, found peroxy radical concentrations in the outflow of convective clouds to be coupled with NO indicating that either NOx and a radical precursor (e.g. formaldehyde, acetone or peroxides) have been simultaneously lifted from lower altitudes or that fresh NO emissions have occurred within uplifted air laden with a radical precursor. Significant O3 production rates of around 1 ppb/h were calculated for these MCS outflows. At the cloud scale, MCSs have contrasting signatures with high O3 production observed for MCSs which had existed for more than 1.5 days as this allows more peroxide formation, and for those MCS which originated south of 10°N where more CO is available for transport to the UT (Ancellet et al., 2009).

The NOx in the UT induces a quasi-persistent large-scale O3 latitudinal gradient with an O3 minimum in the intertropical convergence zone and lightning NOx-related O3 maxima in the southern and the northern Hadley cells (Sauvage et al., 2007; Saunois et al., 2008). However, based on a multi-model study, Barret et al. (2010) demonstrated that global model simulations of lightning NOx (magnitude, altitude and geographical position) (Figure 3) are very sensitive to the convection scheme employed, in particular the detrainment flux levels and intensity. Interestingly, none of the models simulate a NOx maximum over Central Africa as might be expected given the maximum in lightning imaging sensor (LIS) flash frequencies over this region. Simulated O3 enhancements induced by the lightning NOx source are the highest over the northern tropical Atlantic and West Africa.

5. Long-range transport

Surprisingly, the AMMA revealed a persistent influence of fires from the Southern Hemisphere in the mid and lower troposphere of West Africa during the wet season. Import of biomass burning emissions from Central Africa (Liousse et al., 2010) over the southern part of the region was originally proposed by Sauvage et al. (2005) based on MOZAIC data. These incidences of biomass burning import were found to be driven by the Southern Hemisphere African easterly jet activity (Mari et al., 2008) (Figure 4). Signatures of such transport were observed by ozone soundings made in Cotonou, Benin and co-located aircraft measurements with O3 concentrations up to 120 ppbv in the lower troposphere (Thouret et al., 2009) (Figure 4). Significant O3 production rates (7 ppbv/day) have been estimated in middle tropospheric biomass burning plumes transported downwind over the Atlantic Ocean (Real et al., 2010). Global model simulations of the transport of biomass burning emissions show that often transport from Central Africa occurs in the lower troposphere rather than the mid-troposphere when using the ECMWF meteorological analysis (Williams et al., 2010).

Details are in the caption following the image

(Left) Vertical profiles of ozone up to 7 km from the ozone-sounding dataset: August monthly mean and standard deviation (grey), average of the two soundings on 10 and 14/08 (black); from BAe-146 in the region between 5.5°N and 7°N on the 08 and 13/08 (red); and from D-F20 in the region 4–5.5°N on the 13/08 (blue) (adapted from Thouret et al., 2009). (Right) Simulated tracer concentration at 650 hPa on August 15 1200 UTC, originating from the Southern Hemispheric biomass burning emissions during an active phase of the southern African easterly jet (courtesy East Orlandi, adapted from Real et al., 2010)

Central African biomass burning emissions can also be injected periodically into the UT via deep convection over Central Africa and easterly transported in the lower Tropical tropopause layer (TTL) by the Tropical easterly jet (TEJ) (Mari et al., 2008). Slow, but significant, ozone production (1–2 ppbv/day) has been estimated during downwind transport of these air masses around 200 hPa (Real et al., 2010).

Much of the air in the TTL has been advected from the east in the TEJ rather than having been convectively lifted over West Africa. Barret et al. (2008) showed a clear influence of CO-rich air uplifted from Asia (Figure 5). Uplift of clean air over the Indian Ocean as well as transport of air from the lower stratosphere around the Tibetan High also affect trace gas concentrations in the TTL (Law et al., 2010). Interestingly, the TTL also has enhanced levels of non-volatile particles, although the reasons are yet unknown (Borrmann et al., 2009).

Details are in the caption following the image

CO fields, in ppbv, assimilated in the MOCAGE global model, averaged over the period 5–31 July 2006 at 140 hPa (top) and 205 hPa (bottom). Horizontal winds from the ECMWF operational analyses are superimposed as black arrows. White contours at 205 hPa indicate deep convection (OLR contours at 220 W/m2) (adapted from Barret et al., 2008)

6. Challenges for the future

Measurements performed during the AMMA observational periods in 2005 and 2006 gave a first view of the atmospheric composition over West Africa. Several questions remain on the contributions of different components of the vegetation to natural emissions of chemical species and how they change with season and rainfall. How will deforestation and desertification, in a changing climate, modify the chemical emissions? Rapid urbanization and concentration of economic activities in Western Africa's urban centres is an emerging issue for the West African population. The impact of emissions from industry, motor vehicles and households on air quality is a major concern. In situ observations in several cities of West Africa (Dakar, Ouagadougou, Bamako and Lagos) have revealed concentrations of pollutants (black carbon aerosols and NO2) comparable to those observed in Asian megacities (Liousse and Galy-Lacaux, 2010). Current emission inventories estimate only a small contribution of anthropogenic emissions from West Africa, but, based on first evaluations against in situ data, these may be underestimated. The question thus remains concerning the evolution of these emissions especially since Africa is the continent with the highest projected increase in the population in the next few decades.

Acknowledgements

The support of the AMMA project is gratefully acknowledged (see http://onlinelibrary.wiley.com/doi/10.1002/asl.331/full for full acknowledgement).