Persistent organic contaminants in Saharan dust air masses in West Africa,
Cape Verde and the eastern Caribbean
Garrison, V. H., Majewski, M. S., Foreman, W. T., Genualdi, S. A., Mohammed,
A., & Massey Simonich, S. L. (2014). Persistent organic contaminants in Saharan
dust air masses in West Africa, Cape Verde and the eastern Caribbean. Science
of the Total Environment, 468, 530-543. doi:10.1016/j.scitotenv.2013.08.076
10.1016/j.scitotenv.2013.08.076
Elsevier
Version of Record
http://hdl.handle.net/1957/46960
http://cdss.library.oregonstate.edu/sa-termsofuse
Science of the Total Environment 468–469 (2014) 530–543
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Persistent organic contaminants in Saharan dust air masses in West
Africa, Cape Verde and the eastern Caribbean
V.H. Garrison a,⁎, M.S. Majewski b, W.T. Foreman c, S.A. Genualdi d, A. Mohammed e, S.L. Massey Simonich d
a
U.S. Geological Survey, Southeast Ecological Science Center, St. Petersburg, FL 33701, USA
U.S. Geological Survey, California Water Science Center, Sacramento, CA 95819-6129, USA
U.S. Geological Survey, National Water Quality Laboratory, Denver, CO 80225-0585, USA
d
Oregon State University, Department of Environmental and Molecular Toxicology, Corvallis, OR 97331-7301, USA
e
University of the West Indies, Faculty of Science and Technology, St. Augustine, Trinidad and Tobago
b
c
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• Saharan dust storms transport pesticides,
PAHs and PCBs to the eastern Caribbean.
• SOC concentrations 1–3 orders of magnitude greater in Mali than eastern
Caribbean
• SOC atmospheric concentrations were
well below existing occupational guidelines.
• All dust storm samples contained
mixtures of pesticides, PAHs and PCBs.
• Estimated PM10 and PM2.5 concentrations
frequently exceeded USEPA and WHO
limits.
a r t i c l e
i n f o
Article history:
Received 19 July 2013
Received in revised form 23 August 2013
Accepted 23 August 2013
Available online 19 September 2013
Editor: Damia Barcelo
Keywords:
Saharan dust
Long-range transport
Semivolatile organic compounds
West Africa
Caribbean
a b s t r a c t
Anthropogenic semivolatile organic compounds (SOCs) that persist in the environment, bioaccumulate, are toxic at
low concentrations, and undergo long-range atmospheric transport (LRT) were identified and quantified in the atmosphere of a Saharan dust source region (Mali) and during Saharan dust incursions at downwind sites in the eastern Caribbean (U.S. Virgin Islands, Trinidad and Tobago) and Cape Verde. More organochlorine and
organophosphate pesticides (OCPPs), polycyclic aromatic hydrocarbons (PAHs), and polychlorinated biphenyl
(PCB) congeners were detected in the Saharan dust region than at downwind sites. Seven of the 13 OCPPs detected
occurred at all sites: chlordanes, chlorpyrifos, dacthal, dieldrin, endosulfans, hexachlorobenzene (HCB), and
trifluralin. Total SOCs ranged from 1.9–126 ng/m3 (mean = 25 ± 34) at source and 0.05–0.71 ng/m3 (mean =
0.24 ± 0.18) at downwind sites during dust conditions. Most SOC concentrations were 1–3 orders of magnitude
higher in source than downwind sites. A Saharan source was confirmed for sampled air masses at downwind
sites based on dust particle elemental composition and rare earth ratios, atmospheric back trajectory models,
and field observations. SOC concentrations were considerably below existing occupational and/or regulatory limits;
however, few regulatory limits exist for these persistent organic compounds. Long-term effects of chronic exposure
⁎ Corresponding author at: 600 4th Street South, St. Petersburg, FL 33701, U.S.A. Tel.: +1 727 803 8747; fax: +1 727 803 2032.
E-mail address: ginger_garrison@usgs.gov (V.H. Garrison).
0048-9697/$ – see front matter. Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.scitotenv.2013.08.076
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
531
to low concentrations of SOCs are unknown, as are possible additive or synergistic effects of mixtures of SOCs,
biologically active trace metals, and mineral dust particles transported together in Saharan dust air masses.
Published by Elsevier B.V.
1. Introduction
Long-range atmospheric transport (LRT) of semivolatile organic
compounds (SOCs) is one of the primary pathways by which persistent
organic contaminants such as organochlorine and organophosphorous
pesticides (OCPPs), polycyclic aromatic hydrocarbons (PAHs), and
polychlorinated biphenyls (PCBs) are distributed globally (e.g., Peterle,
1969; Risebrough et al., 1968, 1976). The Saharan dust system, largest
in the world (Washington et al., 2003), annually exports an estimated
billion tons of mineral dust from the Sahara–Sahel region of northwest
Africa (e.g., d'Almeida, 1986; Ridley et al., 2012) thousands of kilometers west to the Caribbean and Americas, north to Europe, and east to
Asia, periodically crossing to North America (Creamean et al., 2013;
Hsu et al., 2012; McKendry et al., 2007). This global atmospheric transport system predominately carries eroded mineral soils but biogenic
materials such as pollen (Cariñanos et al., 2004), microorganisms
(e.g., Griffin et al., 2001; Kellogg et al., 2004), particles from fossil fuel
and biomass burning in the Sahel (Formenti et al., 2008; Hand et al.,
2010; Kandler et al., 2011), and insects such as African desert locusts
(Ritchie and Pedgley, 1989) are periodically transported thousands of
kilometers from the dust source area.
The highly heterogeneous Saharan dust air masses (see Reid et al.,
2003 for detailed discussion) vary considerably in size, duration, and
trajectory (Knippertz et al., 2009, 2011; Rajot et al., 2008). Quantities
of dust transported and frequency of episodes vary in response to
surface composition (Chiapello et al., 1997), regional meteorology
(e.g., Engelstaedter and Washington, 2007), location and movement of
the Intertropical Convergence Zone and North Atlantic Oscillation
(Sunnu et al., 2008), and subsidence, advection and convective mixing
of large-scale air masses (Reid et al., 2003). Aerosol composition, concentrations, and sizes vary spatially and temporally within a single
dust episode and among episodes (e.g., Reid et al., 2003) and are a function of geologic (e.g., Ben-Ami et al., 2012; Moreno et al., 2006; Nalli
et al., 2005; Prospero et al., 2002) and geographic origins (Moreno
et al., 2006), meteorological conditions, human activities (e.g., agriculture, land disturbance, transportation, urbanization, and industry)
(e.g., Kandler et al., 2011; Talbot et al., 1986; Tegen and Fung, 1995),
and atmospheric residence time.
Atmospheric SOCs have been well-characterized in some regions but
there is a paucity of information for the Sahara/Sahel and the eastern
Caribbean. Recently established large-scale air monitoring networks
using passive samplers in Africa, Europe and the Americas (e.g., Jaward
et al., 2004c; Klánová et al., 2009; Pozo et al., 2004, 2006; Wong et al.,
2009) provide valuable information on seasonal presence of gas-phase
SOCs and, over time, trends will be identified. However, quantitative
data on atmospheric concentrations of banned and current-use OCPPs,
PAHs, and PCBs in source and remote regions are critical to accurately
quantify exposure, and ultimately, assess risks to ecosystems, humans,
and other organisms. Although transport of eroded mineral dust from
West Africa to the Caribbean and the Americas has been well-studied
(e.g., Delany et al., 1967; Prospero and Nees, 1977; Prospero et al.,
1970, 1981), little is known of transport of other components such as
SOCs. Pioneering work in the mid- to late 1960s in Barbados (Prospero
and Seba, 1972; Risebrough et al., 1968; Seba and Prospero, 1971) detected organochlorine pesticides [dieldrin and dichlorodiphenyltrichloroethane (DDT) and degradates] associated with Saharan dust particles
in the atmosphere but did not consider the gas-phase. Later work in
Barbados sampled both gas and particle-associated phases, finding DDT
and dieldrin levels 10–100 fold higher than previously reported values
for particles (Bidleman et al., 1981) yet an order of magnitude less than
in the Arabian and Red Seas and Persian Gulf (Bidleman and Leonard,
1982) and Gulf of Mexico (Giam et al., 1980). Subsequent to those studies, a number of pesticides widely used in the 1960s have been banned in
much of the world because of their persistence in the environment and
ability to bioaccumulate, biomagnify, and be toxic to humans and other
organisms at low concentrations. Although use of some compounds
has been banned for over three decades, some occur globally in the
atmosphere today.
Pesticides were introduced in West Africa over the past several decades to control agriculture pests and disease vectors [e.g., mosquitoes
(malaria, leishmaniasis), sand flies (onchocerciasis), and tsetse flies
(trypanosomiasis)]. Use has increased substantially, particularly on
export crops such as cotton and rice (van der Valk and Diara, 2000).
Volatilization during and following application and during pesticide
production and repackaging are primary sources of OCPPs to the atmosphere in the dust source region. Other potential sources are fine agricultural soils with sorbed OCPPs and contaminated desert soils. The
Niger River of Mali and Niger is a prime example. Nutrients in flooddeposited sediments and proximity to water make river floodplains
preferred sites to grow crops from garden plots to industrial-scale agriculture. Some OCPPs applied to crops sorb to fine floodplain soils which
in turn can be lifted into the atmosphere by strong winds during the dry
season, entrained in a high altitude transport system such as the Saharan Air Layer (SAL), and transported long distances. Sandy soils on
the edge of the Sahara have been contaminated with DDT and
dieldrin that leaked from containers staged to combat desert locust
outbreaks.
Surface particles in both landscapes are prone to erosion, mobilization and LRT. Wind speed, gustiness (Engelstaedter and Washington,
2007), particle characteristics (size, shape, and composition), and regional meteorology drive the quantities of particles lifted from the surface, altitudes reached, and distances advected from the source area
(e.g., Gillies et al., 1996; Reid et al., 2003). In the classical Saharan dust
transport model, easterly trade winds move the SAL westward off the
African coast, larger particles fall via gravitational settling to the lower
altitudes of the SAL (and eventually the surface), and the dust air mass
becomes enriched in finer particles with time. Generally, the smaller
the particle (and higher the wind speed), the longer it can remain
aloft and the farther it can be transported, such that nearly half of Saharan dust particles are b2.5 μm aerodynamic diameter (PM2.5) when the
SAL reaches the Caribbean (Li-Jones and Prospero, 1998) (Table S.1).
During the past two decades, plastics production, urbanization, and
a rapid increase in motor vehicles (particularly older diesel vehicles
and two-stroke gasoline engines) have resulted in an increased use of
PCBs in electrical equipment and production of PAHs from fossil fuel
combustion. Biomass burning in the countries bordering the Gulf of
Guinea produces PAHs and black carbon that mix with Saharan dust
air masses and undergo LRT to South America and to a lesser extent,
the Caribbean (Rajot et al., 2008). In addition, ubiquitous small, openflame fires throughout West Africa traditionally burned biomass for
cooking fuel and to produce ash for fertilizing crops. Today, plastic
bags, discarded household goods of synthetic materials, and tires are
burned along with biomass (Garrison et al., 2003). PAHs and, potentially, chlorinated dioxins and furans from low temperature combustion,
can rise into the atmosphere and be advected in dust air masses across
oceans to distant continents.
As part of a larger investigation into the effects of Saharan dust incursions on ecosystems and human health, this study was initiated to
(1) identify and quantify atmospheric anthropogenic SOC concentrations and (2) confirm the origin of air masses sampled at downwind
sites off the coast of West Africa, and in the north- and southeastern
Caribbean.
532
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
2. Materials and methods
2.1. Sites
Air was sampled between December 2001 and August 2008
(Table S.2) from (1) a Saharan dust source-region: an urban site in the
Niger River valley in Bamako, Republic of Mali (BKO) and a rural area
on an escarpment above the Niger River near Kati, Mali (KATI) and
(2) locations downwind of Saharan dust regions: offshore of West
Africa [Sal Island, Republic of Cape Verde (CV)] and in the southeastern
[Galera Point, Trinidad (TRIN) and Flagstaff Hill, Tobago (TOB)] and
northeastern Caribbean [East End, St. Croix, U.S. Virgin Islands (VI)]
(Fig. 1). The initial source-region sampling site was moved in 2004
from the Niger River valley (BKO) to the top of the escarpment 8.6 km
to the north (KATI) to minimize urban contamination and for more reliable power. Downwind sampling site locations were carefully selected
to minimize contamination from local sources. Downwind site samples
were collected from the windward coasts/points of the islands and only
when the wind was blowing from the east, off the ocean (see Supplementary material). Availability of sufficient and reliable electricity and
security for equipment and personnel were constraining factors in site
selection at all locations. Despite multiple attempts, samples from the
source-region and downwind sites were not temporally related due to
extreme logistical constraints. Garrison et al. (2011) provides detailed
site information.
2.2. Sampling
A total of 49 samples were collected (Table S.2) from the six sites for
SOC, trace metal, and particle gravimetric analysis. Sampling methods
are described in detail elsewhere (Garrison et al., 2011). Briefly, ambient air was sampled for SOCs and metals in total suspended particles
(TSP) and for SOCs in the gas phase using high volume blower motors.
Particle-associated and gas-phase SOCs were collected on a preconditioned glass fiber filter (GFF) followed by two pre-cleaned polyurethane foam plugs (PUF) housed in Teflon cartridges. Particles for
metal analysis were collected on a preconditioned quartz fiber filter
(QFF) in a nylon and Teflon filter holder. Sampling duration (24–96 h)
and volume (209–822 m3 for SOCs, 231–711 m3 for metals) varied as
a result of atmospheric dust concentrations and filter loading, reliable
and sufficient electricity, and weather (rain, storms). Field blanks
were taken for each type of analysis during each sampling period at
each site and handled identically to samples. All samples were stored
frozen (b−4 °C) until analysis.
2.3. SOC laboratory analysis
Sixty-four targeted historical and current-use OCPPs, six PAHs, and
six PCB congeners (101, 118, 138, 153, 183, and 187) were measured
in the particle and gas-phase components of the air samples. Descriptions of analytical methods used and lists of SOCs analyzed at each laboratory are detailed elsewhere (Garrison et al., 2011).
2.4. Elemental analysis
Total particles were analyzed for 32 elements using inductively
coupled-plasma mass spectrometry (Briggs and Meier, 2002), instrumental neutron activation analysis and/or graphite-furnace atomicabsorption spectrophotometry.
2.5. Data processing
SOC and metal analytical data were field-laboratory blank corrected.
Some OCPP data were pooled prior to data analysis: trans-chlordane,
cis- and trans-nonachlor as “chlordanes”; p,p′-DDT, breakdown products o,p′- and p,p′-DDD and -DDE, and the isomer o,p′-DDT as “DDX”;
and, Endosulfan I, II and sulfate as “endosulfans”. Because ambient
temperatures varied among sites during sampling (21–55 °C), sample
volumes (m3) were corrected to 298 K and 1 atmosphere pressure
using data from local airport meteorology stations near each site (Meteorological Terminal Aviation Routine data).
2.6. Statistical analysis
Random numbers were generated between zero and estimated detection limits and substituted for non-detects in the dataset (Aruga,
1997). To compare SOC profiles among samples, Principal Component
Analysis (PCA) was performed on individual SOC concentrations normalized to total SOCs. PCA and cluster analysis were executed using PRIMER
6® software. Differences in TSP concentrations among source and downwind dust and non-dust conditions were tested using a Student's t-test
(Zar, 2004). Data were further analyzed using non-parametric descriptive statistics (Tables S.3–5) because number of samples varied among
sites and few non-dust-condition samples were collected.
2.7. Air mass origin
Fig. 1. Graphic of sampling sites: ML (Bamako and Kati, Mali, West Africa); CV (Sal Island,
Cape Verde); TT (Galera Point, Trinidad and Flagstaff Hill, Tobago), and VI (St. Croix, U.S.
Virgin Islands).
A weight of evidence approach was used to establish if each air mass
sampled at downwind sites was of Saharan origin by using field observations, analytical data (32 elements), an aerosol forecast/hindcast
model, and an atmospheric transport and dispersion model. All samples
from Mali were considered of Saharan–Sahelian origin. In the field, a
Saharan dust air mass incursion at downwind sites was indicated by
(1) degradation of visibility (Fig. S.1); (2) wind direction between
east-northeast to southeast; (3) presence of reddish-brown particles
on filters; and, (4) dust forecast by the Navy Aerosol Analysis and Prediction System Global Aerosol Model (NAAPS, http://www.nrlmry.
navy.mil/aerosol_web/Docs/globaer_model.html).
Air mass origin of each sample was further investigated based on elemental composition of samples. Specifically, sample iron (Fe) content
(ppm), lanthanum–scandium–thorium (La–Sc–Th) ratios (Muhs et al.,
2007), and enrichment profiles of 30 elements were compared to
Sahara–Sahel dust literature values, source-region samples from this
study, local soils (TOB, VI), and Montserrat volcanic ash (Supplementary
material). Element enrichment factors were calculated against average
upper continental crust elemental composition as in Wedepohl
(1995). Because aluminum (Al) environmental enclosures housed
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
samplers and filters, samples were not analyzed for Al. Back trajectories
of sampled air masses were calculated using the U.S. National Oceanic
and Atmospheric Administration (NOAA) Air Resources Laboratory HYbrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model
(Draxler and Rolph, 2011; Rolph, 2011). Back trajectories for each sampling period were run at 12-h intervals over 10 days, at altitudes both
within (500 m) and at the upper boundary layer limit (1000 m), and
in the SAL (2000 m above ground level because most Saharan dust
transport occurs between 1500 and 7000 m), using CDC1 datasets and
mixing depth.
3. Results and discussion
3.1. Air mass origin
Trace metal composition, elemental enrichment factors, and La–Sc–Th
ratios of total suspended particles (TSP) confirmed that all downwind
samples (CV, TOB, TRIN and VI) collected during Saharan dust-event conditions were of Saharan origin and not local or regional (Figs. 2 and S.2).
La–Sc–Th ratios of BKO and KATI dust samples, Mauritanian and KATI
soils, BKO dry deposition, and downwind dust-incursion samples clustered together, whereas downwind samples collected during non-dust
conditions (air mass origin not Saharan), soils from downwind sample
sites, and Montserrat volcano (Caribbean) ash were distinct from samples
of Saharan origin (Fig. 2). The La–Sc–Th analysis suggested the presence
of Montserrat ash in one VI sample of an air mass that originated over
West Africa (7–9 July 2006; Fig. 2). HYSPLIT back trajectories confirmed
possible regional input from Montserrat volcanic ash in that sample
(Fig. S.3). NAAPS models and HYSPLIT back trajectories provided additional substantiation of a Saharan origin of downwind samples of dustincursions and transport of North American, European or marine influence in non-dust period samples. Regional sources likely contributed
SOCs to two downwind samples: a Saharan dust air mass that moved
over northeastern South America (Fig. S.4) and a non-dust air mass that
traversed the northeastern Caribbean (Fig. S.5) within the boundary
layer and within 24-h of sample collection. In a non-dust period sample,
the movement of boundary layer air from SW Europe directly to the
Sc
10
90
20
80
30
70
40
60
50
50
40
70
30
80
20
90
10
10
20
30
40
50
60
70
Lesser Antilles exemplifies the Atlantic “shortcut” proposed by Gangoiti
et al. (2006) (Fig. S.6).
3.2. Total suspended particulates
TSP concentrations ranged from 93 to 657 μg/m3 (mean = 336 ±
150) in Mali, 10–319 μg/m3 (mean 112 ± 81) at downwind sites during dust conditions, and 28–63 μg/m3 (mean = 40 ± 16) at downwind sites in non-dust conditions (Fig. 3). Atmospheric dust
concentrations were significantly greater in Mali than downwind sites
during dust conditions (p b 0.0005) and in downwind dust than nondust period samples (p b 0.0005). None of the sampling periods coincided with intense dust events at source or downwind sites. TSP concentrations in Mali were 20–200-fold lower than the maximum
(13,735 μg/m3) during an intense dust episode but similar to the
mean background concentrations (575 μg/m3) during the Spring dust
season (April) in Mali reported by Gillies et al. (1996). Downwind site
TSP concentrations during dust incursions were similar to other reports
from the eastern Caribbean (e.g., Perry et al., 1997; Prospero, 1999). Following LRT to the eastern Caribbean, 30–50% of dust particles are PM2.5
(Li-Jones and Prospero, 1999; Perry et al., 1997), a result of gravitational
settling of large particles or aggregates, rainout, and air mass subsidence
removing larger particles during transport (Reid et al., 2003). The U.S.
Environmental Protection Agency (USEPA) and the World Health Organization (WHO), among other health organizations, set regulatory
limits for PM2.5 and particulate matter b 10 μm aerodynamic diameter
(PM10) concentrations, because they are strongly associated with increased mortality and morbidity from cardiorespiratory diseases (e.g.,
Dockery et al., 1993; Wang et al., 2013). PM2.5 and PM10 were estimated
for sampled air masses in this study, using PM2.5 and PM10/TSP ratios
from the literature and an ongoing study in Bamako, Mali (Table S.1):
PM10 reported as 68% of TSP in Mali and the northwest Sahara (Gillies
et al., 1996; Ozer et al., 2007; unpublished data, Garrison et al.); and,
PM2.5 reported as 16–20% of TSP in Mali (d'Almeida and Schütz, 1983;
Gillies et al., 1996; Weinzierl et al., 2009; unpublished data, Garrison
et al.) and 43% of TSP at Caribbean sites (Li-Jones and Prospero, 1998).
All Mali samples (BKO and KATI) exceeded WHO PM10 24-h limit
(50 μg/m3), 74% N USEPA 24-h (150 μg/m3), 84% N WHO PM2.5 24-h
(25 μg/m3), and 74–84% N USEPA 24-h limits (35 μg/m3) (Fig. 3). At
downwind sites in the Caribbean, one half of samples exceeded WHO
and none exceeded USEPA PM10 24-h limits; 67% of Caribbean samples
exceeded WHO and 52% were greater than USEPA PM2.5 24-h limits
(Fig. 3). These findings concur with Perry et al. (1997), Li-Jones and Prospero (1998) and Prospero (1999) that fine particle concentrations in the
majority of Saharan dust incursions in the eastern Caribbean exceed
levels known to be harmful to human health. The situation in the dust
source region is more severe.
3.3. SOCs
60
La
533
80
90
Th
Fig. 2. Ternary diagram of Lanthanum (La)–Scandium (Sc)–Thorium (Th) composition in
sample particles: air samples from dust conditions ( Bamako,
Kati,
Cape Verde,
Trinidad, ▴Tobago, and Virgin Islands); non-dust conditions ( Trinidad and Virgin
Islands); local contamination ( TRINLC);
dry deposition Bamako; soil ( Kati,
Mauritania, Tobago and Virgin Islands); and, Montserrat volcanic ash.
SOCs were detected in all samples [19 source-region and 30 downwind sampling periods of which 26 were from dust and 4 from
non-dust (December and January) periods]. Thirteen banned or
current-use pesticides, six PAHs, and six PCB congeners were identified,
including five of the Dirty Dozen persistent organic pollutants: chlordanes, dieldrin, DDT and breakdown products (DDX), heptachlor and
PCBs. Dust source-region samples contained a larger number of SOCs
and in higher concentrations compared to downwind samples, with
total SOC concentrations one-to-three orders of magnitude higher in
source-region than downwind sites in dust conditions (source: 1.9–
126 ng/m3, mean = 25 ± 34; and downwind: 0.05–0.71 ng/m3,
mean = 0.24 ± 0.18). SOC concentrations were generally an order of
magnitude higher in the source region river valley (BKO, mean =
75 ± 31 ng/m3) than at the escarpment site 8.6 km distant and
170 m higher elevation (KATI, mean = 6.9 ± 3.4 ng/m3). Not only is
greater use of pesticides expected in the urban center with small-scale
534
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
0.7
mg particulates m-3
0.6
0.5
0.4
0.3
0.2
USEPA PM 10 24h
0.1
0
BKO
KATI
CV
TOB
NTRIN
NVI
TRIN
VI
TRIN-LC
WHO PM 10 24h
USEPA PM 2.5 24h
WHO PM 2.5 24h
Fig. 3. Box plot with minimum, maximum, 75% and 25% percentile, arithmetic mean (dashed line) and median (mg/m3) of: measured total suspended particles (no fill); estimated PM10
fraction (gray diagonal lines); and, estimated PM2.5 (gray fill). Bamako (BKO), Kati, Cape Verde (CV), Tobago (TOB), Trinidad (TRIN), Trinidad local contamination (TRIN-LC), Trinidad nondust period (NTRIN), Virgin Islands (VI), and non-dust period Virgin Islands (NVI). U.S.EPA and WHO 24-h limits for PM2.5 and PM10 shown.
agriculture, but an inversion in the river valley (BKO) traps local pollution from numerous small fires, vehicle exhaust, unpaved roads, small
garden plots, and industry. PCA and cluster analysis of individual SOC
concentrations grouped samples generally by site/region: BKO; Kati
and a single TRIN non-dust sample with local contamination (TRINLC); and, downwind sites other than TRIN-LC (Figs. 4, S.7). Despite closer proximity to the source region, CV had lower SOC concentrations
than Caribbean sites (Fig. 5), most likely because the SAL was located
above the boundary layer at that time of year (e.g., Karyampudi et al.,
1999; Reid et al., 2003). CV trace metal composition and La–Sc–Th ratios
confirmed an African origin of particles and by extension, air masses.
Gravitational settling from the SAL would bring dust particles and
sorbed SOCs into the boundary layer at CV (Karyampudi et al., 1999).
The downwind dust-condition sample with greatest SOC concentrations
was collected in TRIN from a Saharan dust air mass (confirmed by La–
Sc–Th ratios, TSP concentrations, visibility, and HYSPLIT back trajectories)
that had briefly transited the coast of northeastern South America within
the boundary layer (Fig. S.4). South America was likely one source of SOCs
10
detected in the sample. One of the four downwind non-dust period samples (TRIN-LC) contained a SOC level 1–2 orders of magnitude greater
than all other downwind samples and within the range found in the
source-region (Fig. 5). TRIN-LC was collected during a non-dust period
with a typical north to northeast wind and coincided with a dump fire
that occurred 42 km to the north (Scarborough, Tobago), inadvertently
providing real-world concentrations from a regional SOC (primarily
PAHs) source. SOC levels in the other downwind non-dust samples
(n = 3) were lower than those from dust conditions but not significantly
(p =0.42) (Fig. 5). Europe was the likely source of elevated SOCs in a VI
non-dust period sample; back trajectories showed the boundary layer
air mass transited directly from Europe to the eastern Caribbean
(Fig. S.6). La–Sc–Th ratios confirmed an origin other than the Sahara
(Fig. 2). LRT of Saharan dust air masses from the Sahara/Sahel and winter
air masses from Europe appear to deliver SOCs to the eastern Caribbean,
confirming one yet contradicting another conclusion of Prospero and
Seba (1972). And, although observed in only one sample in this study, regional SOCs can be transported via Saharan dust air masses that briefly
transit northeastern South America prior to entering the eastern Caribbean. SOC concentrations in all samples were considerably below existing
occupational and/or regulatory limits, although few guidelines on concentrations in air exist for the SOCs detected in this study.
BKO
5
PC2
KATI
CV
0
TOB
TRIN
NTRIN
-5
TRIN-LC
VI
NVI
-10
-20
-15
-10
-5
0
5
PC1
101
102
103
log SOCs
Fig. 4. PCA biplot of individual SOCs in samples. All Bamako sites cluster; Kati sites
cluster; TRIN-LC
clusters with Kati sites; TR050702
clusters with but differs from
other downwind sites; and, all downwind sites (except TRIN-LC) cluster. PC1 accounts
for 51.6% and PC2 accounts for 13.3% of variation (64.9% total).
104
105
106
pg m-3
Fig. 5. Box plots showing minimum, maximum, 75% and 25% percentile, arithmetic mean
(dashed line) and median sum of SOCs (log pg/m3) by site. Random number between zero
and estimated detection limit substituted for no detection.
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
3.3.1. Banned and current-use pesticides
Seven of the 13 banned and current-use pesticides and their isomers
and degradation products identified in air samples in this study [chlordanes, chlorpyrifos, dacthal, dieldrin, endosulfans, hexachlorobenzene
(HCB), and trifluralin] were found at all sites and chlorpyrifos, endosulfans, and HCB were detected in most samples (Fig. 6). Nine [those listed
above plus DDX, and γ-hexachlorocyclohexane (γ-HCH)], occurred at
both source and downwind sites. Source region samples contained
one-to-multiple orders of magnitude higher concentrations of chlordanes, chlorpyrifos, dacthal, dieldrin, DDX, endosulfans, and trifluralin
than downwind sites (Fig. 6). Heptachlor and α-HCH were detected
only at downwind sites (TOB; TRIN, TOB and non-dust VI, respectively),
whereas diazinon and profenofos were identified only in the source region (BKO). Endosulfans, HCB and chlorpyrifos were the most abundant
OCPPs detected in this study and with the exception of endosulfans, differed from two investigations of atmospheric SOCs in the northwestern
U.S., an area impacted by LRT from Asia as well as local and regional
sources: HCB, endosulfans, and α-HCH (Genualdi et al., 2009); and
dacthal, endosulfans, and γ-HCH (Primbs et al., 2008).
3.3.1.1. Source and downwind sites. Chlordanes, carcinogenic and mutagenic organochlorine insecticides, occurred in up to 3-fold higher concentrations in urban BKO (19–86 pg/m3, mean = 48 ± 28, median = 28)
than KATI (not detected (nd) — 4.1 pg/m3, mean = 1.5 ± 1.3, median = 1.2) and downwind sites (nd — 27 pg/m3, mean = 2.5 ± 4.9, median = 1.4; Fig. 6). BKO appears to be a source of chlordane for regional
and LRT. trans-Chlordane and trans-nonachlor were detected in most
samples, whereas cis-chlordane was not detected and cis-nonachlor was
detected only at the VI site (Garrison et al., 2011). Chlordane levels at
downwind and KATI sites were within the lower range of concentrations
reported from the eastern Atlantic (Gioia et al., 2012), similar to an industrial site in Turkey (Bozlaker et al., 2009), and lower by 2–3 orders of magnitude in residences in three U.S. cities (Offenberg et al., 2004). A single
high downwind concentration (27 pg/m3 in TRIN) was similar to some
remote sites, including Barbados in the 1970s when chlordane was still
in use (e.g., Bidleman et al., 1981; Bidleman and Leonard, 1982) and within the range reported in the eastern north Atlantic off West Africa (Gioia
et al., 2012), pointing to a possible local or regional source. Back trajectories showed the air mass primarily within the boundary layer over the Atlantic, with short periods over the Sahara–Sahel within the mixed layer,
briefly transiting northeastern South America within the boundary layer
(Fig. S.4), and over the Atlantic above the boundary layer. The Sahara,
South America and/or local re-emission or use all are possible chlordane
sources.
The MONET passive sampling network detected little chlordane at
Mali stations, including Bamako (Klánová et al., 2009). However, their
estimates in Ghana (Adu-Kumi et al., 2012) bracketed BKO concentrations in this study. In contrast, Pozo et al. (2009) did not detect chlordanes in Ghana in 2005. Overall, chlordanes in BKO appear to be in
line with other urban centers with current use and re-emission and
would be expected to be a source of particle-associated (from soil
resuspension) and volatilized chlordanes for regional and LRT. Downwind site concentrations were generally at background levels with
occasional increases from long-distance and likely, regional transport.
Chlorpyrifos, a current-use organophosphate insecticide that binds
strongly to soil (Verschueren, 1996), also volatilizes and exists primarily
in the gas phase in air (Majewski et al., 2008). It is used on cotton,
peanuts, and vegetable crops and against fleas, cockroaches, mosquitos,
termites and other pests around dwellings, and on pets. Chlorpyrifos
was one of the most abundant and frequently detected pesticides (75%
of samples) and occurred at all sites except TOB. Chlorpyrifos in BKO
(urban and close to cotton and maize growing areas) reached nearly
30,000 pg/m3, with mean concentrations (9960 ± 11,000, median =
3500) 200-fold greater than the neighboring rural and non-agricultural
KATI site (nd — 220 pg/m3, mean = 44 ± 63, median = 22), and 3
orders of magnitude greater than in downwind samples during dust
535
(nd — 8.3 pg/m3, mean = 2.1 ± 1.9, median = 2.2) and non-dust conditions (nd — 20 pg/m3; mean = 5.7 ± 9.4, median = 1.3) (Fig. 6). A
single non-dust period sample (TRIN-LC) that captured local contamination from a nearby garbage dump fire contained 20 pg/m3 chlorpyrifos
likely volatilized by the fire. Chlorpyrifos levels in BKO were similar to
those reported in a use area in California (LeNoir et al., 1999) and orders
of magnitude greater than those reported at an industrial site in Turkey
(Bozlaker et al., 2009). Air concentrations in the mid-continental U.S.A.
in the 1990s (a high use area; Majewski et al., 1998, 2008) and in both agricultural and non-use areas in Quebec (Sadiki and Poissant, 2008) were
greater than KATI and less than BKO. KATI levels were similar to locations
impacted by regional transport: remote tropical and subtropical mountains of Brazil (estimates from passive samplers; Meire et al., 2012) and
far less than in Costa Rica close to local use (Gouin et al., 2008b). Although
chlorpyrifos occurred in 70% of downwind samples, its concentrations
were low as would be expected from LRT.
The organochlorine insecticide DDT is licensed for use against disease
transmitting insects and some crop pests in the Sahara/Sahel. DDT and/
or its breakdown products (DDX) were found in 80% of source region
but only 3% of downwind samples (p,p′-DDE in a single TRIN dustcondition sample). Concentrations in BKO samples (2450–3800 pg/m3;
mean = 2900 ± 590, median = 2600) were 2–3 orders of magnitude
greater than at the KATI escarpment site (1.6–1010 pg/m3; mean =
110 ± 260, median = 42; Fig. 6). Higher concentrations in the river valley likely are due to a concentrated population using DDT on crops and
against malarial mosquitoes, volatilization from a pesticide formulation
and repackaging facility in BKO, trapping by an inversion layer, and/or
greater use during the time period sampled (2001–2003) in BKO than
KATI (2004–2008). DDE/DDT ratios generally increased from 0.6
(2001) to 1.0 (2003) in BKO and from 0.5 (2004) to 2.6 (2006) in KATI
with no DDT detected after 2006 (Garrison et al., 2011). This could be
year-to-year variability or indicative of decreasing DDT use. Mixtures
of detected DDX components changed over time. All six DDT isomers
and degradates were present from 2001 to 2004 in the source region
(p,p′-DDE N p,p′-DDT N o,p′-DDT N p,p′-DDD N o,p′-DDD N o,p′-DDE),
whereas only three (p,p′-DDE N p,p′-DDT N o, p′-DDE) were detected
from 2006 to 2008. DDX concentrations in BKO were considerably
higher than most values reported since DDT was banned (e.g., Klánová
et al., 2009; Pozo et al., 2009), except in areas where DDT was in use at
the time (e.g., Alegria et al., 2000, 2008; Karlsson et al., 2000; Weber
and Montone, 1990). KATI values were: higher than those reported in
the eastern Atlantic off West Africa (Gioia et al., 2012), as would be
expected of a continental source to windward; consistent with those reported from some oceanic sites and non-agricultural areas (Jaward et al.,
2004a; Klánová et al., 2009; Pozo et al., 2009); and, lower than some
urban and agricultural regions (Alegria et al., 2000, 2008; Karlsson
et al., 2000; Weber and Montone, 1990). The TRIN sample p,p′-DDE concentration (0.7 pg/m3) was 1–3 orders of magnitude lower than KATI
and BKO, 2–20 fold lower than estimates in Ghana (Adu-Kumi et al.,
2012; Pozo et al., 2009), but similar to values from the eastern Atlantic
off West Africa that were attributed to LRT (Gioia et al., 2012). This pattern is consistent with LRT from West Africa. p,p′-DDT and p,p′-DDE
estimates from particles only in Barbados in the late 1960s when DDT
was still in wide use (mean = 0.1 pg/m3; Prospero and Seba, 1972;
Seba and Prospero, 1971) were lower than the single TRIN detection in
this study. Measuring both gas and particle associated phases,
Bidleman et al. (1981) reported up to 5 pg/m3 of p,p′-DDT and -DDE
and DDT/DDE ratios indicative of current use. In this study, o,p′-DDT
was detected only in the source region (Table S.3), indicating recent
use. All source region sampling occurred during the dry (dust) season
when pesticide use is relatively low, whereas the single downwind sample that contained DDX (TRIN p,p′-DDE) was collected during the West
African rainy season (July) when DDT application against mosquitoes
and other disease vectors occurs. Seba and Prospero (1971) suggested
that DDT detected in Saharan dust incursions in the eastern Caribbean
originated in Europe or North America, not Africa, at that time and was
536
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
A
total chlordanes
chlorpyrifos
dacthal
BKO
5/5
KATI
12/14
10/14
CV
3/5
3/5
TOB
4/4
0/4
1/4
TRIN
6/10
10/10
2/10
NTRIN
1/1
1/1
1/1
TRIN-LC
1/1
1/1
0/1
VI
7/7
6/7
3/7
NVI
2/2
1/2
1/2
10-3 101
101
103
5/5
105
10-2
100
102
log pg
B
BKO
DDX
1/5
1/14
1/5
104
10-3 10-1
dieldrin
5/5
5/5
CV
0/5
1/5
4/5
TOB
0/4
3/4
4/4
1/10
3/10
8/10
NTRIN
0/1
1/1
0/1
TRIN-LC
0/1
0/1
1/1
VI
0/7
1/7
7/7
NVI
0/2
0/2
2/2
10-3
100
104
103
C
BKO
14/14
100
log pg
104
103
100
104
m-3
HCB
trifluralin
3/5
0/5
1/5
KATI 14/14
1/14
3/14
CV
5/5
0/5
1/5
TOB
4/4
4/4
1/4
5/10
3/10
TRIN 10/10
NTRIN
1/1
0/1
0/1
TRIN-LC
1/1
0/1
0/1
VI
7/7
5/7
0/7
NVI
2/2
1/2
1/2
10-1
100
101
105
5/5
10/14
TRIN
103
sum endosulfans
4/14
KATI
101
m-3
102
10-2
100
103
10-3
100
102
log pg m-3
Fig. 6. Box plots showing minimum, maximum, 75% and 25% percentile, arithmetic mean and median pesticide concentrations (pg/m3) in source region and downwind by site. Random number between zero and estimated detection limit substituted for no detection. Number of samples in which analyte was detected/number of samples analyzed shown to the right of y-axis.
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
scavenged from the atmosphere over the Atlantic by Saharan dust. However, in this study, a Sahara/Sahel source is suspected based on air mass
back trajectories, the presence of active SAL dust transport from Africa
(Mali and Mauritania) to the Caribbean (including TRIN) during the
sampling period, La–Sc–Th ratios of dust, and, detection of p,p′-DDE
only and at a low concentration. Current use of DDT in West Africa
makes the region a likely source for LRT.
Dieldrin, an organochlorine insecticide, is currently used against
disease vectors such as malarial mosquitoes and other insect pests
(e.g., termites, locusts, and cotton pests) in the Sahel. It was detected
in all BKO samples and one third of KATI and downwind site samples.
BKO concentrations (91–1800 pg/m3, mean = 1100 ± 840, median =1700) exceeded KATI levels (nd — 47 pg/m3, mean = 10 ±
16, median = 2.0) by N 2 orders of magnitude and downwind sites
(nd — 21 pg/m3, mean = 2.6 ± 4.2, median = 0.6) by 2–3 orders of
magnitude. Levels in BKO were similar to those reported from an inland
agricultural area of Belize (Alegria et al., 2000) but greater than midcontinental U.S.A. sites (Majewski et al., 2008), and were multiple
orders of magnitude higher than estimated from the MONET passive
sampling network in Mali, Senegal (Klánová et al., 2009) and Ghana
(Adu-Kumi et al., 2012). The KATI dieldrin concentrations exceeded
the highest MONET passive sampler estimates in Mali (urban Bamako;
Klánová et al., 2009) and Ghana (Adu-Kumi et al., 2012) but were within those reported from Lake Malawi, a site with small-scale agriculture
(Karlsson et al., 2000). Data from our study do not show a decrease in
dieldrin in source site samples over time, indicating that the difference
between BKO and KATI levels is likely a function of greater use in urbanagricultural BKO. The comparatively low levels reported by MONET
could be an artifact of mass per volume estimates from passive samplers
in conditions of extreme temperatures and humidity. Concentrations at
downwind sites appeared to be at background levels, similar to remote
sites when dieldrin was more widely in use (e.g., Atlas and Giam, 1981;
Bidleman et al., 1981; Bidleman and Leonard, 1982) and at background
sites near urban centers today (Adu-Kumi et al., 2012). Soils contaminated by dieldrin spills at staging sites for locust-control in northern
Mali are another source of dieldrin. West Africa appears to be a source
for regional and LRT of dieldrin.
Endosulfan (ENDO), an organochlorine insecticide and acaricide, was
the most abundant pesticide detected, as has been reported by investigations in other regions of the world (e.g., reviewed in Weber et al.,
2010). Total ENDO concentrations followed patterns observed for
other SOCs: BKO (11,200–15,900 pg/m3, mean = 13,100 ± 1800,
median = 12,400) exceeded KATI (19–460 pg/m3, mean = 180 ±
140, median = 140) which exceeded downwind sites (nd — 45 pg/m3,
mean = 8.6 ± 9.5, median = 4.7; Fig. 6). Concentrations at downwind
sites during dust incursions on average were twice those in non-dust conditions. BKO concentrations were within recent values reported from use
areas in Central America (e.g., Daly et al., 2007; Wong et al., 2009) and
higher than estimates from passive samplers in Ghana, Senegal and
Mali (Klánová et al., 2009; Pozo et al., 2009), Bermuda and the Canary
Islands (Shunthirasingham et al., 2010). KATI levels were similar to
areas affected by regional transport (e.g., Hayward et al., 2010; Pozo
et al., 2004; Shunthirasingham et al., 2010), lower than other West
African countries (Klánová et al., 2009; although these are estimates
from passive samplers), and greater than values reported from a site
(Lake Malawi; Karlsson et al., 2000) with small-scale agriculture.
ENDO I, II and sulfate occurred in all source region samples at 80, 16,
and 16%, respectively, of downwind samples. The ENDO I:II ratios in
source samples (1.0–5.6, mean = 2.7) indicate local use of technical
ENDO (ratios 2:1 to 7:3) (Herrmann, 2002). Lower concentrations and
ENDO I:II ratios at downwind sites (3.3–26, mean = 14, with no
ENDO II detection in 83% of samples) were consistent with isomerization of II to I in an aged air mass, as would be expected with LRT
(e.g., Weber et al., 2010). Of 20 pesticides applied in the cottongrowing area of Mali from 1997 to 2007, ENDO was the most widely
used, followed by profenofos, and chlorpyrifos (Dem et al., 2007).
537
ENDO II was the most commonly detected pesticide in soil in four cotton
growing regions in Mali (Dem et al., 2007) likely due to greater volatilization of ENDO I than II from soils (Antonious et al., 1998). Other than
HCB which was identified in all samples, ENDO and chlorpyrifos were
the most frequently detected OCPPs at downwind sites and reflect
OCPP use in the dust source region. West Africa continues to be a source
of ENDO available for regional and LRT.
HCB, one of the most abundant chlorinated hydrocarbons in the
atmosphere in the late 1980s (Duce et al., 1991), was detected in 96%
of samples at generally low concentrations (nd — 113 pg/m3,
mean = 8.6 ± 16, median = 5.6) similar to reported global background levels (e.g., Genualdi et al., 2009; Gioia et al., 2012; Jaward
et al, 2004c; Klánová et al., 2009; Shen et al., 2005). Two KATI samples
at higher concentrations (26 and 110 pg/m3) indicate local use.
Technical hexachlorocyclohexane (HCH), no longer in use, contained
five HCH isomers: γ-HCH (lindane), the active isomer (10–15%); α-HCH
(60–70%); and β-, δ-, and ε-HCH. Today, lindane, composed primarily of
γ-HCH, is banned from use in 52 countries, restricted in many more
since 2009, and is currently licensed for use in Mali and other Saharan/
Sahelian countries. It is used primarily to treat seeds such as sorghum
(the major cultivated crop in Mali) against insect pests and to control
external parasites of humans, livestock, and pets, although it also has
been used to preserve fish in Mali (Tingle, 2008). γ-HCH was detected more frequently in downwind dust-condition (50%) than sourceregion samples (5%), with TOB concentrations 6–20 fold higher than
VI, TRIN, and the single source region (KATI) samples (Fig. 6). TOB γHCH values (30–52 pg/m3, mean = 41 ± 9.1, median = 42) were
an order of magnitude lower than those reported in some areas
with local use (e.g., Dvorská et al., 2009; Hoff et al., 1992; Karlsson
et al., 2000) but similar to others with current use [an agricultural
area in Belize in the 1990s (Alegria et al., 2000), urban and rural
areas in Canada (Gouin et al., 2008a), and an urban area in Chile
(Pozo et al., 2004)]. Local use in TOB appears probable, despite the
relatively low concentrations. Concentrations from all sites in this
study (nd — 52 pg/m3) were within the range reported in the eastern
Atlantic off West Africa (b 3.6–100 pg/m3) (Jaward et al., 2004a), where
West Africa is considered the probable source. TRIN and VI concentrations were similar to those reported from remote sites (e.g., Atlas and
Giam, 1981; Atlas and Schauffler, 1990; Bidleman and Leonard, 1982;
Genualdi et al., 2009; Jaward et al., 2004c) and lower than areas with
current agricultural and urban use (e.g., Alegria et al., 2000; Hoff et al.,
1992; Jaward et al., 2004c; Karlsson et al., 2000), further strengthening
the case for a remote source. Zhang et al. (2008) reported trans-Atlantic
atmospheric transport of lindane from Western Africa/Western Europe
to the Caribbean, southern and eastern U.S. Considering the documented current use of γ-HCH in Mali (Diarra, 1997), the concentrations and
low frequency of detection in BKO and KATI samples might be a reflection of the time of year sampling occurred. All source region sampling
occurred from Dec–May, during the primary dust season in Mali but
prior to the planting (and rainy) season when most pesticide application occurs (June–Sept). All TOB and VI and half of TRIN dust air mass
sample collection occurred during the source region rainy season
(June to October), the period of maximum pesticide use. LRT from
West Africa is one source of γ-HCH at downwind sites, with probable
local use in TOB.
Atmospheric concentrations of α-HCH have declined since 2000 (Li
et al., 2003) in response to decreased use of technical HCH and lindane,
and changes in lindane production that reduced the amount of
α-isomer. Surprisingly, α-HCH was detected in TOB only and in low
concentrations (b1.4 pg/m3) similar to those reported in the eastern
Atlantic by Gioia et al. (2012), but lower than those reported in the eastern Atlantic off West Africa by Jaward et al. (2004c) and in the Indian
Ocean and Western Pacific (Gioia et al., 2012), and at the lowest concentrations reported from urban and suburban Ghana (Adu-Kumi et al.,
2012), where lindane remains in use. α-HCH detected in Tobago may
have come from photolysis of γ-HCH or re-emission from soil treated
538
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
decades ago (reviewed in Walker et al., 1999). The low α-:γ-ratio
(b 0.04, n = 2) in TOB is indicative of LRT (α- and γ-isomers) and local
use (γ-HCH). However, identification of a probable local source has
been elusive. Lindane use in TRIN-TOB is limited to treating head lice
and for external parasites of household pets (UNEP/POPS/POPRC.6/13/
Annex II); however, no dwellings or roads were windward of the TOB
sampling site. Although bananas and other fruits (not grain crops) are
grown on the steep hillside that slopes to the sea windward of the
sampling site, the windward topography, land use, and adjacent land
use do not conform with a local source of lindane. Although considered
an unlikely source, a controlled burn of nearby grass-covered land
(downwind of the sampling site) occurred prior to sampling and may
have volatilized historic residues of one or both isomers in the biomass
or soil (Genualdi et al., 2009). Low α-:γ-ratios (b0.04), low-tomoderate concentrations, back trajectories, and time of year point to
LRT from West Africa to downwind sites, although local use of lindane
cannot completely be ruled out in TOB.
Toxicity of dacthal, a current-use organochlorine herbicide, arises primarily from the highly toxic impurities, HCB and dioxins (e.g., http://
pmep.cce.cornell.edu/profiles/herb-growthreg/dalapon-ethephon/dcpa/
herb-prof-dcpa.html). BKO concentrations (nd — 4.0 pg/m3, mean =
1.0 ± 1.7, median = nd) exceeded KATI (nd — 0.2 pg/m3, mean =
0.1 ± 0.3, median = nd) and downwind sites (nd — 0.3 pg/m3,
mean = 0.2 ± 0.1, median = nd) by an order of magnitude. Highest
concentrations in BKO were lower than those reported from locations
with local or regional use: the Canary Islands (Gioia et al., 2011); northwest U.S. (Genualdi et al., 2009; Primbs et al., 2008); and, the Mississippi
River Valley (Majewski et al., 1998). BKO concentrations were similar to
those reported from remote sites: during non-growing season at a forested site subject to regional transport (Daly et al., 2007); and, subject to
long-distance and regional transport (Genualdi et al., 2009; Killin et al.,
2004). Dacthal is not approved for use in the Sahel although illegal use
is considered likely. Background-level concentrations at all sites suggest
that West Africa is not a source of dacthal for LRT.
Trifluralin concentrations were low in all sites and samples
(Fig. 6) (BKO, nd — 2.4 pg/m3, mean = 1.4 ± 0.9, median = nd;
KATI, nd — 1.1 pg/m3, mean = 0.3 ± 0.3, median = nd; and downwind sites, nd — 2.9 pg/m3, mean = 0.2 ± 0.5, median = nd).
Levels were similar to a remote site in Ontario, Canada (Hoff et al.,
1992) and orders of magnitude lower than other published values
(e.g., Daly et al., 2007; Gioia et al., 2011; LeNoir et al., 1999;
Majewski et al., 1998, 2008; Sadiki and Poissant, 2008). Trifluralin,
a current-use herbicide commonly used on grass and crops such as
fruits, vegetables, cotton and soybeans was banned in Europe in
2008 and is not on the list of pesticides used in the Sahel. Despite
the low levels of trifluralin detected, local or regional use cannot be
ruled out as sources.
3.3.1.2. Source region only. Diazinon and profenofos are organophosphate insecticides used on field crops, fruits and vegetables, and around
residences. Only BKO, KATI, and TRIN samples collected in 2001–2004
were analyzed for profenofos. Diazinon and profenofos were detected
only in BKO samples from 2001 to 2003 (nd — 2.6 ng/m3, mean =
0.8 ± 1.1, median = 0.2; and nd — 1.7 ng/m3, mean = 0.9 ± 0.8,
median = 1.3, respectively). Diazinon in BKO air was 3 orders of magnitude greater than that reported in some other use areas (LeNoir
et al., 1999; Majewski et al., 1998, 2008), yet similar to the median for
Jackson, MS, U.S.A. (Majewski et al., 2008) and some reported maxima
(U.S. Environmental Protection Agency, 2000), indicating that the
Sahel could be a source of diazinon for transport. BKO does not appear
to be a source of profenofos considering the relatively low concentrations [3 orders of magnitude lower concentrations than Thailand
(Jaipieam et al., 2009)]. Susceptibility to photodecomposition processes,
coupled with considerations of physical properties (Mackay et al.,
1997), makes substantive LRT of diazinon and profenofos unlikely.
3.3.1.3. Downwind sites only. Heptachlor, a legacy organochlorine
insecticide, was identified only in downwind site samples. No degradation product (the epoxide) was detected. Concentrations in
TOB (nd — 22 pg/m3, mean = 13 ± 9.1, median = 15) and TRIN
(nd — 54 pg/m3, mean =13 ± 17, median = 7.9), frequency of detection, and lack of epoxide indicate a local source (e.g., Gioia et al.,
2011; Hoff et al., 1992; Karlsson et al., 2000; Klánová et al., 2009) or
revolatilization from soil with historic use. Heptachlor appears on the
list of pesticides used in the Sahel but was not detected in BKO or KATI,
likely a result of season sampled. Regional transport from the British
Virgin Islands or St. Martin appears the most likely source of the single
occurrence (4 pg/m3) in a VI non-dust sample (Fig. S.5).
3.3.2. PAHs
PAHs enter the atmosphere from: incomplete combustion of fossil
fuels, wood, waste, and other substances; volatilization from soils (anaerobic degradation of organic matter); the petroleum refining process,
flares, and spills; and other sources. Most PAHs have been shown to be
sorbed primarily on the respirable fraction of particles (Manichini,
1992). The sum of six PAHs (Σ6PAH) detected in West Africa in this
study (Fig. S.8) was similar to levels reported in the eastern Atlantic
off the West African coast (Jaward et al., 2004c; Nizzetto et al, 2008)
and concentrations at downwind sites were similar to Σ10PAHs in the
relatively pristine southeastern Atlantic (Nizzetto et al., 2008). Σ6PAHs
in this study were 1–3 orders of magnitude higher in source region
samples and a single non-dust sample from Trinidad (TRIN-LC) than
in downwind samples (Fig. S.8). This is consistent with the continual
burning of biomass and garbage throughout the dust source region
and the rapid increase in use of two-stroke engines. In the TRIN-LC sample, a dump fire 42 km to the north coincided with non-dust condition
sampling. The high Σ6PAHs concentrations in this sample were within
the range of KATI and BKO values and provided PAH concentrations
from a substantiated regional source (Fig. S.8). Σ6PAH levels were not
significantly different among the other downwind site samples. CV
levels were marginally higher than other downwind sites overall
(TOB, TRIN and VI) but lower than source sites with local/regional
sources (BKO, KATI and TRIN-LC), conforming with a LRT model.
Phenanthrene, fluoranthene and pyrene, were detected at all sites
and were the most frequently detected PAHs (Fig. 7), as reported in
other investigations (e.g., Halsall et al., 1994; Jaward et al., 2004b).
Nizzetto et al. (2008) found phenanthrene, fluoranthene, and pyrene
dominated atmospheric PAHs in the eastern Atlantic off West Africa
but not Europe and attributed that to oil extraction (and refineries, flares,
spills, etc.), biomass burning, and natural sources in West Africa.
Rodríguez et al. (2011) concluded that oil refineries and other Northwest
African industries were probable sources of industrial pollutants in the
SAL. Phenanthrene levels reported in the eastern Atlantic off West
Africa (mean = 1.6 ng/m3; Nizzetto et al., 2008) were: lower than
sites with local and regional sources [BKO (mean = 13 ± 10 ng/m3),
KATI (mean = 3.2 ± 1.6 ng/m3) and TRIN-LC (3.8 ng/m3)]; and higher
than sites affected by LRT [downwind dust-condition sites in this study
(mean = 0.1 ± 0.2 ng/m3)]. Benzo [a] pyrene is strongly carcinogenic
(U.S. Centers for Disease Control, 1995) and its concentrations were
higher in source sites than the few downwind samples (TOB and
TRIN) in which it was identified (Fig. 7). Anthracene, a combustion
product with a short half-life, was detected in 11 of 19 source samples but in only one downwind sample (12 pg/m3) at a concentration close to the method detection limit (Fig. 7). Anthraquinone, a
photodegradate of anthracene, also has industrial (dye manufacturing,
pulp bleaching) and other sources (http://toxnet.nlm.nih.gov/cgi-bin/
sis/search/a?dbs+hsdb:@term+@DOCNO+2074); it occurred in source
region samples only. Cloth and leather dying occurs at small- and
medium-scale enterprises throughout the BKO area.
Individual PAH concentrations at downwind sites (excepting TRINLC) in this study (Fig. 7) were consistently within or below the low
range of values reported from remote sites (Atlas and Schauffler,
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
A
anthracene
BKO
anthraquinone
5/5
benzo[a]pyrene
5/5
3/5
KATI
2/14
1/14
12/14
CV
0/5
0/5
0/5
TOB
0/4
0/4
1/4
TRIN
1/10
0/10
1/10
NTRIN
0/1
0/1
0/1
TRIN-LC
0/1
0/1
1/1
VI
0/7
0/7
0/7
NVI
0/2
0/2
0/2
10-3
10-1
101
103
10-2
100
102
log pg
B
fluoranthene
104
10-2
100
phenanthrene
5/5
5/5
5/5
14/14
14/14
14/14
CV
5/5
5/5
5/5
TOB
4/4
4/4
4/4
TRIN
9/10
6/10
10/10
NTRIN
1/1
1/1
1/1
TRIN-LC
1/1
1/1
1/1
VI
7/7
7/7
7/7
NVI
2/2
2/2
2/2
105
10-3 10-1 101
log pg
104
pyrene
KATI
103
102
m-3
BKO
10-3 10-1 101
539
103
105 10-3 10-1 101
103
105
m-3
Fig. 7. Box plots showing minimum, maximum, 75% and 25% percentile, arithmetic mean and median individual PAH concentrations (pg/m3) by site. Random number between zero and
estimated detection limit substituted for no detection. Number of samples in which analyte was detected/number of samples analyzed shown to the right of y-axis.
1990; Genualdi et al., 2009) and considerably lower than other published studies (e.g., Del Vento and Dachs, 2007; Jaward et al., 2004b;
Manichini, 1992; Masclet et al., 1988). Source site PAH (anthracene,
benzo [a] pyrene, fluoranthene, phenanthrene) values were higher
than most reports from remote sites (Atlas and Schauffler, 1990;
Genualdi et al., 2009; Jaward et al., 2004b) and within the range of or
higher than a remote Mediterranean location (Masclet et al., 1988), an
urban area in Florida, U.S. (Poor et al., 2004), and West Africa-impacted
sites in the Atlantic (Del Vento and Dachs, 2007; Jaward et al., 2004b).
Overall, PAH concentrations near sources (BKO, KATI, TRIN-LC) were
similar to those reported from other PAH-source locations, whereas
concentrations in the eastern Caribbean were similar to remote sites
and regions impacted by LRT. Based on the findings from a study of seawater/air partitioning of PAHs in the eastern Atlantic, atmospheric
transport from primary sources drives atmospheric PAH concentrations
in the Atlantic, with little to no influence from volatilization from seawater (Nizzetto et al., 2008). The pattern of PAH presence/absence
and concentrations, rare earth “tracers” in atmospheric dust showing a
West African origin of sampled air masses, meteorological conditions,
and air mass back trajectories in this study indicate LRT from the Sahara/Sahel as the primary source of PAHs during Saharan dust conditions
in the eastern Caribbean.
3.3.3. PCBs
More PCB congeners (101, 118, 138, 153, 183, 187) were detected and
occurred in higher concentrations in dust source locations than at downwind sites (Fig. 8), similar to patterns seen for OCPPs and PAHs. Congener
concentrations decreased from BKO N KATI N TRIN N CV, TOB, VI. Higher
concentrations in TRIN than other downwind sites suggest an upwind regional PCB source — possibly from ships passing to windward or offshore
gas platforms to the SSE. Source region sample Σ6PCBs (nd — 52 pg/m3,
mean = 14 ± 12, median = 10) were nearly an order of magnitude
higher than in downwind dust (nd — 11 pg/m3, mean = 2.3 ± 238, median = nd) and non-dust samples (nd — 6.0 pg/m3, mean = 2.3 ± 2.5,
median = nd). The downwind sample with local contamination from a
dump fire (TRIN-LC) contained elevated PCB concentrations (5.7 pg/m3)
that were similar to dust source sites (Fig. 8). PCB 183 was detected in
the source-region only and co-planar PCB (118), a product of open
540
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
A
PCB 101
PCB 118
5/5
BKO
PCB 138
1/5
3/5
1/14
KATI
11/14
12/14
CV
1/5
0/5
0/5
TOB
0/4
1/4
3/4
TRIN
2/10
8/10
1/10
NTRIN
1/1
1/1
0/1
TRIN-LC
1/1
1/1
0/1
VI
2/7
1/7
0/7
NVI
0/2
0/2
0/2
10-2
B
100
102
10-1
101
log pg m-3
PCB 153
10-2
PCB 183
100
102
PCB 187
BKO
1/5
1/5
1/5
KATI
12/14
10/14
12/14
CV
0/5
0/5
0/5
TOB
2/4
0/4
3/4
1/10
0/10
0/10
0/1
0/1
TRIN
NTRIN
1/1
TRIN-LC
0/1
0/1
0/1
VI
0/7
0/7
0/7
NVI
0/2
0/2
0/2
10-3
10-1
101
10-3
10-1
101 102 10-3
10-1
101 102
log pg m-3
Fig. 8. Box plots showing minimum, maximum, 75% and 25% percentile, arithmetic mean and median PCB concentrations (pg/m3) by site. Random number between zero and estimated
detection limit substituted for no detection. Number of samples in which analyte was detected/number of samples analyzed shown to the right of y-axis.
burning, was detected at all sites except CV. Lack of PCBs in CV samples
could be a consequence of the short period sampled and/or the altitude
of the SAL.
Individual congener concentrations along an eastern Atlantic transect off Africa (Gioia et al., 2012) were within the range observed in
this study in West Africa (BKO and KATI) and at downwind sites, as
would be expected with transport from West Africa to the west. BKO
and KATI congener concentrations were similar to those reported in
West Africa by Gioia et al. (2011), higher than along a N–S eastern Atlantic transect (Gioia et al., 2008, 2012), but lower than another N–S
eastern Atlantic transect (Jaward et al., 2004c). The spatial pattern of
PCB concentrations seen in our study and compared to other investigations supports LRT of PCBs from the Sahara/Sahel to the Caribbean. PCB
concentrations in downwind dust-condition samples were similar to
those reported from the remote SE Atlantic (Gioia et al., 2008) and
lower than other remote sites (e.g., Dachs et al., 1999).
Detection of PCB 101 only at CV and at a lower concentration than in
other studies (Gioia et al., 2008, 2011; Schreitmüller and Ballschmiter,
1994) may be a function of time of year, altitude of the SAL and mixing
depth during sampling, and/or differences in air mass pollution load
when sampling occurred. Considerably lower PCB concentrations in
the Caribbean than West Africa are to be expected from dilution during
LRT. West Africa has previously been identified as a source of PCBs available for LRT (Gioia et al., 2011). Sources in West Africa include electrical
equipment such as transformers, plastics manufacturing (such as in
Bamako, Mali; Diarra, 1997), and open burning of waste, including contaminated soils and vegetation (Eckhardt et al., 2007). PCB reemission
from the ocean has been suggested as a secondary source to the atmosphere (Jaward et al., 2004a) and could contribute to background levels
in the insular Caribbean.
3.4. Implications for Saharan dust source regions and the eastern Caribbean
What does the presence of atmospheric SOCs that are toxic to organisms (biotoxic) and inhalable particles mean for humans and other
organisms in the Sahara/Sahel and in the eastern Caribbean during
Saharan dust incursions? To begin to address this important but unstudied question, we used the concentration data to estimate quantities of
SOCs that could be inhaled by a person in one day (ng/d) at each sampling location. This provides an approximation of SOCs that could enter
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
the body through the nose and eventually reach the lungs (gas-phase
and sorbed on PM10) or stomach (PM10–100 trapped in mucus)(Johnson
and Vincent, 2003). Assumptions are: (a) reported SOC concentrations
(both gaseous and particle-associated SOCs) are inhaled, even though
some fraction is sorbed to particles too large to be inhaled (N PM100);
(b) the amount of SOCs sorbed to NPM100 fraction is small because
most detected SOCs were likely in the gas-phase and no particles
N80 μm were reported in Mali by Gillies et al. (1996), indicating a low
NPM100 fraction in the source region and downwind; (c) estimates
represent inhalable maximums during the light-to-moderate dust conditions sampled; and, (d) a conservative normal respiration rate for a resting adult (12 breaths/min) and an average tidal volume of 0.5 L/breath
were used (des Jardins, 2008). Estimates of PM10 from TSP were based
on literature values and an ongoing study (Table S.1).
We estimated that a scientist sampling in the dust source region
could have inhaled 17–1100 ng/d (mean = 200 ± 300) SOCs and
0.4–6 ng/d (mean = 2 ± 2) at downwind sites during dust conditions.
During the periods sampled, estimated inhalable SOCs were an order of
magnitude greater at the urban dust source region site (BKO, 440–
1100 ng/d, mean = 650 ± 300) than at a more rural location 8.6 km
distant and 170 m higher in elevation (KATI, 17–110 ng/d, mean =
60 ± 30). For individual analytes, estimated inhalation was 1–3 orders
of magnitude greater in source than downwind sites.
As discussed in Section 3.2, USEPA and WHO PM10 24-h limits were
exceeded 74–100% of days sampled in the source region and 0–50% of
days in the eastern Caribbean during dust conditions, respectively.
USEPA and WHO PM2.5 24-h limits were exceeded 74–84% of days sampled in the source region and 52–67% of days in the eastern Caribbean
during dust conditions, respectively. Elevated PM10 and PM2.5 concentrations present recognized adverse effects on human cardiorespiratory
health (e.g., Dockery et al., 1993; Pope et al., 1995; Wang et al., 2013).
People (and other organisms) in the Sahara/Sahel are exposed to
considerably higher concentrations of inhalable particles as well as
biotoxic synthetic compounds than at downwind sites following LRT.
Presently, no information exists on the effects of SOCs or fine particles
and associated metals on health in the source region (de Longueville
et al., 2010) or at downwind sites in the eastern Caribbean and
elsewhere. Investigations to elucidate the effects of chronic exposure
to atmospheric SOCs, fine particles, and mixtures on human health are
needed.
4. Conclusions
The Sahara/Sahel region of Africa is a source of OCPPs, PAHs, and
PCBs exported via the SAL to the eastern Caribbean, a seemingly remote
region considered to have unimpaired air quality. Most SOC concentrations were orders of magnitude higher in the dust source region than at
downwind sites during dust incursions but also were orders of magnitude lower than the few existing regulatory limits. Inhalation limits do
not exist for many of the SOCs detected in this study and no guidelines
exist for mixtures. Mixtures of SOCs and respirable particles containing
bioactive and biotoxic metals were present in all dust-condition
samples. PM2.5 and PM10 concentrations in sampled dust air masses
often exceeded existing air quality standards above which there are
recognized adverse effects on human cardiorespiratory health. Little is
known about toxicity or effects of SOC mixtures or in combination
with bioactive/biotoxic metals on organism health and ultimately on
ecosystems. Are there SOC–SOC and SOC–metal interactions? If so, are
they synergistic, additive or antagonistic? Are there adverse health effects from: chronic, low level exposure to multiple biotoxic compounds;
or, periodic exposure to biotoxic compounds and fine particles that can
be inhaled into the lungs? There is an urgent need to elucidate the longterm effects and mechanisms by which real-world mixtures of biotoxic
and bioactive synthetic compounds and mineral dust in the atmosphere
impact human health and ecosystems.
541
Conflict of interest statement
The authors declare that they have no competing financial or other
conflicting interests.
Acknowledgments
The authors thank the following individuals and agencies who were
essential to the project in Africa, the Caribbean, and the U.S.: S. Coulibaly,
M. Dandara, O. Konipo, D. Maiga, B. Nadio, M. Ranneberger, R. Smith,
American International School, Ministry of Communication (ORTM)
and staff at emetteur Kati, Ministry of Geology and Mines, U.S. Embassy
Bamako (Mali); P.J.P. Gomes, J. Pimiento, E. Santos Suarez, Institute of
Meteorology and Geophysics, International Airport and Security
Authority (Sal Island, Cape Verde); J. Agard, S. Mahabir, L. Thomas, Environmental Management Authority, Maritime Services, Tobago House
Assembly, University of the West Indies (Trinidad and Tobago); R.
Berey, H. Tonnemacher, R. Van Heckman (St. Croix, U.S. Virgin Islands);
and P. Lamothe, J. Schrlau, S. Skaates, and G. Wilson (sample analysis).
The NOAA Air Resources Laboratory (ARL) is acknowledged for the provision of the HYSPLIT transport and dispersion model website (http://
ready.arl.noaa.gov) used in this publication. Deep thanks to the Friends
of Virgin Islands National Park, National Aeronautic and Space Administration, and the U.S. Geological Survey for providing project financial
support and the U.S. Embassy Bamako for in-kind support. Funding
sponsors supported performance of the research and article preparation
only and had no involvement in: study design, sample collection, analysis and interpretation of data; report preparation or submission for publication. Any use of trade names is for descriptive purposes only and
does not imply endorsement by the U.S. Government.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.scitotenv.2013.08.076.
References
Adu-Kumi S, Kareš R, Literák J, Borůková J, Yeboah P, Carboo D, et al. Levels and seasonal
variations of organochlorine pesticides in urban and rural background air of southern
Ghana. Environ Sci Pollut Res 2012;19:1963–70.
Alegria HA, Bidleman TF, Shaw TJ. Organochlorine pesticides in ambient air of Belize,
Central America. Environ Sci Technol 2000;34:1953–8.
Alegria HA, Wong F, Jantunen LM, Bidleman TF, Figueroa MS, Bouchot GG, et al. Organochlorine pesticides and PCBs in air of southern Mexico (2002–2004). Atmos Environ
2008;42:8810–8.
Antonious GF, Byers ME, Snyder-Conn E. Residues and fate of endosulfan on fieldgrown
pepper and tomato. Pestic Sci 1998;54:61–7.
Aruga R. Treatment of responses below the detection limit: some current techniques
compared by Factor Analysis on environmental data. Anal Chim Acta 1997;354:
255–62.
Atlas E, Giam CS. Global transport of organic pollutants: ambient concentrations in the
remote marine atmosphere. Science 1981;211:163–5.
Atlas EL, Schauffler S. Concentration and variation of trace organic compounds in the
North Pacific atmosphere. In: Kurtz DA, editor. Long range transport of pesticides.
Chelsea, MI: Lewis Publishers; 1990. p. 161–84.
Ben-Ami Y, Koren I, Altaratz O, Kostinski A, Lehahn Y. Discernible rhythm in the spatio/
temporal distributions of transatlantic dust. Atmos Chem Phys 2012;12:2253–62.
Bidleman TF, Leonard R. Aerial transport of pesticides over the Northern Indian ocean and
adjacent seas. Atmos Environ 1982;16:1099–107.
Bidleman TF, Christensen EJ, Billings WN, Leonard R. Atmospheric transport of organochlorines in the North Atlantic gyre. J Mar Res 1981;39:443–64.
Bozlaker A, Muezzinoglu A, Odabas M. Processes affecting the movement of organochlorine pesticides (OCPs) between soil and air in an industrial site in Turkey.
Chemosphere 2009;77:1168–76.
Briggs PH, Meier AL. The determination of 42 elements in geological materials by ICP–MS.
In: Taggert JE, editor. Analytical methods for chemical analysis of geologic and other
materials, USGS Open file report 02–223. Denver, Colorado, USA: US Geol Survey;
2002. p. 1–14.
Cariñanos P, Galan C, Alcazar P, Dominguez E. Analysis of the particles transported with
dust-clouds reaching Cordoba, southwestern Spain. Arch Environ Contam Toxicol
2004;46:141–6.
Chiapello I, Bergametti GL, Chatenet B, Bousque P, Dulac F, Santos-Suares E. Origins of
African dust transport over northeastern tropical Atlantic. J Geophys Res 1997;102:
13,701–9.
542
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
Creamean JM, Suski KJ, Rosenfeld D, Cazorla A, DeMott PJ, Sullivan RC, et al. Dust and
biological aerosols from the Sahara and Asia influence precipitation in the Western
U.S. Science 2013;339:1572–8.
d'Almeida GA. A model for Saharan dust transport. J Climate Appl Meteorol 1986;25:
903–16.
Dachs J, Bayona JM, Ittekkot V, Albaiges J. Monsoon-driven vertical fluxes of organic
pollutants in the Western Arabian Sea. Environ Sci Technol 1999;33:3949–56.
d'Almeida GA, Schütz L. Number, mass and volume distributions of mineral aerosol and
soils of the Sahara. J Climate Appl Meteorol 1983;22:233–43.
Daly GL, Lei YD, Teixeira C, Muir DCG, Castillo LE, Jantunen LMM, et al. Organochlorine
pesticides in the soils and atmosphere of Costa Rica. Environ Sci Technol 2007;41:
1124–30.
de Longueville F, Hountondji YC, Henry S, Ozer P. What do we know about effects of desert dust on air quality and human health in West Africa compared to other regions?
Sci Total Environ 2010;409:1–8.
Del Vento S, Dachs J. Atmospheric occurrence and deposition of Polycyclic Aromatic
Hydrocarbons in the northeast tropical and subtropical Atlantic Ocean. Environ Sci
Technol 2007;41:5608–13.
Delany AC, Delany AC, Parkin DW, Griffin JJ, Goldberg ED, Reimann BEF. Airborne dust
collected at Barbados. Geochim Cosmochim Acta 1967;31:885–909.
Dem SB, Cobb JM, Mullins DE. Pesticide residues in soil and water from four cotton growing areas of Mali, West Africa. J Agric Food Environ Sci 2007;1. [http://www.
scientificjournals.org/journals2007/articles/1062.htm, Accessed 11 July 2013].
des Jardins TR. Cardiopulmonary anatomy & physiology. 5th ed. NY, USA: DELMAR
CENGAGE Learning; 2008.
Diarra A. Pollution in Mali. United National Environment Program Proceedings of the Subregional Awareness Raising Workshop on Persistent Organic Pollutants (POPs),
Bamako, Mali, 15–18 December 1997; 1997. [http://www.chem.unep.ch/pops/
POPs_Inc/proceedings/bamako/eng/Mali.html, Accessed 11 July 2013].
Dockery DW, Pope III CA, Xu X. An association between air pollution and mortality in six
US cities. N Engl J Med 1993;329:1753–9.
Draxler RR, Rolph GD. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory)
Model access via NOAA ARL READY Website (http://ready.arl.noaa.gov/HYSPLIT.php).
Silver Spring, MD: NOAA Air Resources Laboratory; 2011 [Accessed 11 July 2013].
Duce RA, Liss PS, Merrill JT, Atlas EL, Buat-Menard P, Hicks BB, et al. The atmospheric input
of trace species to the world ocean. Global Biogeochem Cycles 1991;5:193–259.
Dvorská A, Lammel G, Holoubek I. Recent trends of persistent organic pollutants in air in
central Europe — air monitoring in combination with air mass trajectory statistics as a
tool to study the effectivity of regional chemical policy. Atmos Environ 2009;43:1280–7.
Eckhardt S, Breivik K, Mano S, Stohl A. Record high peaks in PCB concentrations in the
Arctic atmosphere due to longrange transport of biomass burning emissions. Atmos
Chem Phys 2007;7:6229–54. [Disc].
Engelstaedter S, Washington R. Temporal controls on global dust emissions: the role of
surface gustiness. Geophys Res Lett 2007;34:L15805. http://dx.doi.org/10.1029/
2007GL029971.
Formenti P, Rajot JL, Desboeufs K, Caquineau S, Chevaillier S, Nava S, et al. Regional variability of the composition of mineral dust from western Africa: results from the
AMMA SOP0/DABEX and DODO field campaigns. J Geophys Res 2008;113:D00C13.
Gangoiti G, Alonso L, Navazo M, García JA, Millán MM. North African soil dust and
European pollution transport to America during the warm season: hidden links
shown by a passive tracer simulation. J Geophys Res 2006;111:D10109.
Garrison VH, Shinn EA, Foreman WT, Griffin DW, Holmes CW, Kellogg CA, et al. African
and Asian dust: from desert soils to coral reefs. Bioscience 2003;53:469–80.
Garrison VH, Foreman WT, Genualdi SA, Majewski MS, Mohammed A, Simonich SM.
Concentrations of semivolatile organic compounds associated with African dust air
masses in Mali, Cape Verde, Trinidad and Tobago, and the U.S. Virgin Islands,
2001–2008. US Geological Survey Data SeriesReston, VA, USA: US Geol Survey;
2011. [http://pubs.usgs.gov/ds/571, Accessed 15 July 2013].
Genualdi S, Killin RK, Woods J, Wilson G, Schmedding D, Massey Simonich SL.
Trans-Pacific and regional atmospheric transport of polycyclic aromatic hydrocarbons and pesticides in biomass burning emissions to Western North America.
Environ Sci Technol 2009;43:1061–6.
Giam CS, Atlaas E, Chan HS, Neff GS. Phthalate esters, PCBs, and DDT residues in the Gulf
of Mexico atmosphere. Atmos Environ 1980;14:65–9.
Gillies J, Nickling WG, McTainsh G. Dust concentrations and particle-size characteristics of
an intense dust haze event: Inland Delta Region, Mali, West Africa. Atmos Environ
1996;30:1081–90.
Gioia R, Nizzetto L, Lohmann R, Dachs J, Temme C, Jones KC. Polychlorinated biphenyls
(PCBs) in air and seawater of the Atlantic Ocean: sources, trends and processes.
Environ Sci Technol 2008;42:1416–22.
Gioia R, Eckhardt S, Breivik K, Jaward FM, Prieto A, Nizzetto L, et al. Evidence for major
emissions of PCBs in the West African Region. Environ Sci Technol 2011;45:1349–55.
Gioia R, Li J, Schuster J, Zhang Y, Zhang G, Li X, et al. Factors affecting the occurrence and
transport of atmospheric organochlorines in the China Sea and the Northern Indian
and South East Atlantic Oceans. Environ Sci Technol 2012;46:10012–21.
Gouin T, Shoeib M, Harner T. Atmospheric concentrations of current-use pesticides across
south-central Ontario using monthly-resolved passive air samplers. Atmos Environ
2008a;42:8096–104.
Gouin T, Wania F, Ruepert C, Castillo LE. Field testing passive air samplers for current use
pesticides in a tropical environment. Environ Sci Technol 2008b;42:6625–30.
Griffin DW, Garrison VH, Herman JR, Shinn EA. African desert dust in the Caribbean atmosphere: microbiology and public health. Aerobiologia 2001;17:203–13.
Halsall CJ, Coleman PJ, Davis BJ, Burnett V, Waterhouse KS, Harding-Jones P, et al. Polycyclic aromatic hydrocarbons in U.K. urban air. Environ Sci Technol 1994;28:2380–6.
Hand VL, Capes G, Vaughan DJ, Formenti P, Haywood JM, Coe H. Evidence of internal
mixing of African dust and biomass burning particles by individual particle analysis
using electron beam techniques. J Geophys Res — Atmos 2010;115. http://dx.doi.org/
10.1029/2009jd012938.
Hayward SJ, Gouin T, Wania F. Comparison of four active and passive sampling techniques
for pesticides in air. Environ Sci Technol 2010;44:3410–6.
Herrmann M. Preliminary risk profile of endosulfan. Berlin Germany: Umweltbundesamt;
2002.
Hoff RM, Muir DCG, Grift NP. 1992. Annual cycle of polychlorinated biphenyls and
organohalogen pesticides in air in southern Ontario. 1. Air concentration data. Environ Sci Technol 1992;26:266–75.
Hsu S-C, Huh C-A, Lin C-Y, Chen WN, Mahowald NM, Liu S-C, et al. Dust transport from
non-East Asian sources to the North Pacific. Geophys Res Lett 2012;39:L12804.
http://dx.doi.org/10.1029/2012GL051962.
Jaipieam S, Visuthismajarn P, Siriwong W, Borjan M, Robson MG. Inhalation exposure of
organophosphate pesticides by vegetable growers in the Bang-Rieng Subdistrict in
Thailand. J Environ Public Health 2009. http://dx.doi.org/10.1155/2009/452373.
Jaward MF, Barber JL, Booij K, Dachs J, Lohmann R, Jones KC. Evidence for dynamic air–
water coupling and cycling of persistent organic pollutants over open Atlantic
Ocean. Environ Sci Technol 2004a;38:2617–25.
Jaward MF, Barber JL, Booij K, Jones KC. Spatial distribution of atmospheric PAHs
and PCNs along a north–south Atlantic transect. Environ Pollut 2004b;132:
173–81.
Jaward MF, Farrar NJ, Harner T, Sweetman AJ, Jones KC. Passive air sampling of PCBs,
PBDEs, and organochlorine pesticides across Europe. Environ Sci Technol 2004c;38:
34–41.
Johnson D, Vincent J. Sampling and sizing of airborne particles. In: DeNardi SR, editor. The
occupational environment: its evaluation, control, and management. Fairfax, VA:
American Industrial Hygiene Assoc; 2003.
Kandler K, Lieke K, Benker N, Emmel C, Küpper M, Müller-Ebert D, et al. Electron microscopy of particles collected at Praia, Cape Verde, during the Saharan Mineral Dust
Experiment: particle chemistry, shape, mixing state and complex refractive index.
Tellus B 2011;63:475–96.
Karlsson H, Muir DCG, Teixiera CF, Burniston DA, Strachan WMJ, Hecky RE, et al. Persistent chlorinated pesticides in air, water, and precipitation from the Lake Malawi
area, Southern Africa. Environ Sci Technol 2000;34:4490–5.
Karyampudi VM, Palm SP, Reagan JA, Fang H, Grant WB, Hoff RM, et al. Validation of the
Saharan dust plume conceptual model using lidar, Meteosat and ECMWF data. Bull
Am Meteorol Soc 1999;80:1045–75.
Kellogg CA, Griffin DW, Garrison VH, Peak KK, Royall N, Smith RR, et al. Characterization
of aerosolized bacteria and fungi from desert dust events in Mali, West Africa.
Aerobiologia 2004;20:99–110.
Killin RK, Simonich SL, Jaffee DA, DeForest CL, Wilson GR. Transpacific and regional
atmospheric transport of anthropogenic semivolatile organic compounds to
Cheeka Peak Observatory during the Spring of 2002. J Geophys Res 2004;109:
D23S15. http://dx.doi.org/10.1029/2003JD004386.
Klánová J, Cupr P, Holoubek I, Boruvkova J, Pribylova P, Kares R, et al. Monitoring of persistent organic pollutants in Africa. Part 1: Passive air sampling across the continent
in 2008. J Environ Monit 2009;11:1952–63.
Knippertz P, Ansmann A, Althausen D, Müller D, Tesche M, Bierwirth E, et al. Dust
mobilization and transport in the northern Sahara during SAMUM 2006 — a meteorological overview. Tellus B 2009;61:12–31.
Knippertz P, Tesche M, Heinold B, Kandler K, Toledano C, Esselbor M. Dust mobilization
and aerosol transport from West Africa to Cape Verde — a meteorological overview
of SAMUM-2. Tellus Ser B Chem Phys Meteorol 2011;63:430–47.
LeNoir JS, McConnell LL, Fellers GM, Cahill TM, Seiber JN. Summertime transport of
current-use pesticides from California's Central Valley to the Sierra Nevada Mountain
Range, USA. Environ Toxicol Chem 1999;18:2715–22.
Li Y-F, Scholtz MT, van Heyst BJ. Global gridded emission inventories of βhexachlorocyclohexane. Environ Sci Technol 2003;37:3493–8.
Li-Jones X, Prospero JM. Variations in the size distribution of non-sea-salt sulfate aerosol
in the marine boundary layer at Barbados: impact of African dust. J Geophys Res —
Atmos 1998;103:16073–84.
Mackay D, Shiu WY, Ma KC. Illustrated handbook of physical–chemical properties and environmental fate for organic chemicals–Pesticide chemicals. Boca Raton, FL: Lewis
Publishers; 1997.
Majewski MS, Foreman WT, Goolsby DA, Nakagaki N. Airborne pesticide residues along
the Mississippi River. Environ Sci Technol 1998;32:3689–98.
Majewski MS, Foreman WT, Coupe RH, Goolsby DA, Wiebe FW. Pesticides in air and rainwater in the midcontinental United States, 1995—methods and data. Reston, VA: US
Geol Survey; 2008.
Manichini E. Urban air pollution by polycyclic aromatic hydrocarbons: levels and sources
of variability. Sci Total Environ 1992;106:109–35.
Masclet P, Pistikopoulos P, Beyne S, Mouvier G. Long-range transport and gas/particle distribution of polycyclic aromatic hydrocarbons at a remote site in the Mediterranean
Sea. Atmos Environ 1988;22:639–50.
McKendry IG, Strawbridge KB, O'Neill NT, Macdonald AM, Liu PSK, Leaitch WR, et al.
Trans-Pacific transport of Saharan dust to western North America: a case study.
J Geophys Res — Atmos 2007;112:D01103. http://dx.doi.org/10.1029/2006jd007129.
Meire RO, Lee SC, Yao Y, Targino AC, Torres JPM, Harner T. Seasonal and altitudinal variations of legacy and current-use pesticides in the Brazilian tropical and subtropical
mountains. Atmos Environ 2012;59:108–16.
Moreno T, Querol X, Castillo S, Alastuey A, Cuevas E, Herrmann L, et al. Geochemical
variations in aeolian mineral particles from the Sahara–Sahel Dust Corridor.
Chemosphere 2006;65:261–70.
Muhs DR, Budahn JR, Prospero JM, Carey SN. Geochemical evidence for African dust
inputs to soils of western Atlantic islands: Barbados, the Bahamas, and Florida.
J Geophys Res 2007;112. http://dx.doi.org/10.1029/2005JF000445.
V.H. Garrison et al. / Science of the Total Environment 468–469 (2014) 530–543
Nalli NR, Clemente-Colón P, Morris V, Joseph E, Szczodrak M, Minnett PJ, et al. Profile observations of the Saharan air layer during AEROSE 2004. Geophys Res Lett 2005;32:
L05815.
Nizzetto L, Lohmann R, Gioia R, Jahnke A, Temme C, Dachs J, et al. PAHs in air and seawater along north–south transect: trends, processes and possible sources. Environ Sci
Technol 2008;42:1580–5.
Offenberg JH, Naumova YY, Turpin BJ, Eisenreich SJ, Morandi MT, Stock T, et al. Chlordanes
in the indoor and outdoor air of three U.S. cities. Environ Sci Technol 2004;8:2760–8.
Ozer P, Laghdaf MBOM, Lemine SOM, Gassani J. Estimation of air quality degradation due
to Saharan dust at Nouakchott, Mauritania, from horizontal visibility data. Water Air
Soil Pollut 2007;178:79–87.
Perry KD, Cahill T, Eldred R, Dutcher DD. Long-range transport of North African dust to the
eastern United States. J Geophys Res 1997;102:11,225–38.
Peterle TJ. DDT in Antarctic snow. Nature 1969;224:620.
Poor N, Tremblay R, Kay H, Bhethanabotla V, Swartz E, Luther M, et al. Atmospheric
concentrations and dry deposition rates of polycyclic aromatic hydrocarbons
(PAHs) for Tampa Bay, Florida, USA. Atmos Environ 2004;38:6005–15.
Pope CA, Thun MJ, Namboodiri MM, Dockery DW, Evans J, Speizer FE, et al. Particulate air
pollution as a predictor of mortality in a prospective study of U.S. adults. Am J Respir
Crit Care Med 1995;151:669–74.
Pozo K, Harner T, Shoeib M, Urrutia R, Barra R, Parra O, et al. Passive-sampler derived air
concentrations of persistent organic pollutants on a North–South transect in Chile.
Environ Sci Technol 2004;38:6529–37.
Pozo K, Harner T, Wania F, Muir DCG, Jones KC, Barrie LA. Toward a global network for
persistent organic pollutants in air: results from the GAPS study. Environ Sci Technol
2006;40:4867–73.
Pozo K, Harner T, Lee SC, Wania F, Muir DCG, Jones KC. Seasonally resolved concentrations
of persistent organic pollutants in the global atmosphere from the first year of the
GAPS Study. Environ Sci Technol 2009;43:796–803.
Primbs T, Wilson G, Schmedding D, Higginbotham C, Massey Simonich S. Influence of
Asian and Western United States agricultural areas and fires on the atmospheric
transport of pesticides in the Western United States. Environ Sci Technol 2008;42:
6519–25.
Prospero JM. Long-term measurements of the transport of African mineral dust to the
southeastern United States: implications for regional air quality. J Geophys Res
1999;104:15917–27.
Prospero JM, Nees RT. Dust concentration in the atmosphere of the Equatorial North
Atlantic: possible relationship to the Sahelian drought. Science 1977;196:1196–8.
Prospero JM, Seba DB. Some additional measurements of pesticides in the lower atmosphere of the northern equatorial Atlantic Ocean. Atmos Environ 1972;6:363–4.
Prospero JM, Bonatti E, Schubert C, Carlson TN. Dust in the Caribbean atmosphere traced
to an African dust storm. Earth Planet Sci Lett 1970;9:287–93.
Prospero JM, Glaccum RA, Nees RT. Atmospheric transport of soil dust from Africa to
South America. Nature 1981;289:570–2.
Prospero JM, Ginoux P, Torres O, Nicholson SE, Gill TE. Environmental characterisation of
global sources of atmospheric soil dust identified with the Nimbus 7 total ozone
mapping spectrometer (TOMS) absorbing aerosol product. Rev Geophys 2002;40:
1002. http://dx.doi.org/10.1029/2000RG000095.
Rajot JL, Formenti P, Alfaro S, Desboeufs K, Chevaillier S, Chatenet B, et al. AMMA dust experiment: an overview of measurements performed during the dry season special
observation period (SOP0) at the Banizoumbou (Niger) supersite. J Geophys Res
2008;113:D00C14.
Reid JS, Kinney JE, Westphal DL, Holben BN, Welton EJ, Tsay S-C, et al. Analysis of measurements of Saharan dust by airborne and ground-based remote sensing methods
during the Puerto Rico Dust Experiment (PRIDE). J Geophys Res — Atmos
2003;108. http://dx.doi.org/10.1029/2002jd002493.
Ridley DA, Heald CL, Ford B. North African dust export and deposition: a satellite and
model perspective. J Geophys Res — Atmos 2012;117:D02202. http://dx.doi.org/
10.1029/2011JD016794.
Risebrough RW, Huggett RJ, Griffin JJ, Goldberg ED. Pesticides: transatlantic movements
in the northeast trades. Science 1968;159:1233–6.
Risebrough RW, Walker II W, Schmidt TT, De Lappe BW, Connors CW. Transfer of chlorinated biphenyls to Antarctica. Nature 1976;264:738–9.
Ritchie J, Pedgley D. Desert Locust cross the Atlantic. Antenna 1989;13:10–2.
Rodríguez S, Alastuey A, Alonso-Pérez S, Querol X, Cuevas E, Abreu-Afonso J, et al. Transport of desert dust mixed with North African industrial pollutants in the subtropical
Saharan Air Layer. Atmos Chem Phys 2011;11:6663–85.
543
Rolph GD. Real-time Environmental Applications and Display sYstem (READY) Website
(http://ready.arl.noaa.gov). Silver Spring, MD: NOAA Air Resources Laboratory;
2011 [Accessed 16 July 2013].
Sadiki M, Poissant L. Atmospheric concentrations and gas-particle partitions of pesticides:
comparisons between measured and gas-particle partitioning models from source
and receptor sites. Atmos Environ 2008;42:8288–99.
Schreitmüller J, Ballschmiter K. Levels of polychlorinated biphenyls in the lower troposphere of the North- and South-Atlantic Ocean. Fresenius J Anal Chem 1994;348:
226–39.
Seba DB, Prospero JM. Pesticides in the lower atmosphere of the northern equatorial
Atlantic Ocean. Atmos Environ 1971;5:1043–50.
Shen L, Wania F, Lei YD, Teixeira C, Muir DCG, Bidleman TF. Atmospheric distribution and
long-range transport behavior of organochlorine pesticides in North America.
Environ Sci Technol 2005;39:409–20.
Shunthirasingham C, Oyiliagu CE, Cao X, Gouin T, Wania F, Lee S-C. Spatial and temporal
pattern of pesticides in the global atmosphere. J Environ Monit 2010;12:1650–7.
Sunnu A, Afeti G, Resch F. A long-term experimental study of the Saharan dust presence in
West Africa. Atmos Res 2008;87:13–26.
Talbot RW, Harris RC, Browell EV, Gregory GL, Sebacher DI, Beck SM. Distribution and geochemistry of aerosols in the tropical north Atlantic troposphere: relationship to
Saharan dust. J Geophys Res 1986;91:5173–82.
Tegen I, Fung I. Contribution to the atmospheric mineral aerosol load from land surface
modification. J Geophys Res 1995;100:18,707–26.
Tingle CCD. Pesticides & poverty: implementing chemical conventions for safe and just development. Pesticide ecotoxicology: documenting and communicating environmental
impacts. Final report on follow up mini projects. UK: Pesticide Action Network; 2008
[http://www.thenrgroup.net/theme/PAN-ecotox/pdf/eco-tox_mini%20projects_
technical_report.pdf, Accessed 16 July 2013].
U.S. Centers for Disease Control Agency for Toxic Substances, Disease Registry (ATSDR).
Toxicological profile for Polyaromatic Hydrocarbons (PAHs) 1995. Atlanta, GA, USA:
USEPA; 1995 [http://www.atsdr.cdc.gov/toxprofiles/tp69.html, Accessed 16 July
2013].
U.S. Environmental Protection Agency. Environmental risk assessment for diazinon. U.S.
Environmental Protection Agency, Office of Prevention, Pesticides and Toxic
Substances, Office of Pesticide Programs. Washington, DC: U.S. Government Printing
Office; 2000.
van der Valk H, Diarra A. Pesticide use and management in the African Sahel — an overview. In: van der Valk H, editor. Les Pesticides au Sahel Utilisation, impact
et alternatives, 4–5. Sahelian Studies and Research, Institut du Sahel; 2000. p. 14–29.
Verschueren K. Handbook of environmental data on organic chemicals. 3rd ed. New York,
NY: Van Nostrand Reinhold; 1996.
Walker K, Vallero DA, Lewis RG. Factors influencing the distribution of Lindane and other
hexachlorocyclohexanes in the environment. Environ Sci Technol 1999;33:4373–8.
Wang X, Chen R, Meng X, Geng F, Wang C, Kan H. Associations between fine particle,
coarse particle, black carbon and hospital visits in a Chinese city. Sci Total Environ
2013;458–460:1–6.
Washington RM, Todd C, Middleton N, Goudie AS. Dust-storm source areas determined
by the total ozone monitoring spectrometer and surface observations. Ann Assoc
Am Geogr 2003;93:297–313.
Weber RR, Montone RC. Distribution of organochlorines in the atmosphere of the south
Atlantic and Antarctic Oceans. In: Kurtz DA, editor. Long range transport of pesticides.
Chelsea, Michigan: Lewis Publishers; 1990. p. 185–97.
Weber J, Halsall CJ, Muir D, Teixeira C, Small J, Solomon K, et al. Endosulfan, a global pesticide: a review of its fate in the environment and occurrence in the Arctic. Sci Total
Environ 2010;408:2966–84.
Wedepohl KH. The composition of the continental crust. Geochim Cosmochim Acta
1995;59:1217–32.
Weinzierl B, Petzold A, Esselborn M, Wirth M, Rasp K, Kandler K, et al. Airborne measurements of dust layer properties, particle size distribution and mixing state of Saharan
dust during SAMUM 2006. Tellus B 2009;61:96–117.
Wong F, Alegria HA, Alvarado V, Angeles F, Galarza AA, Bandala ER, et al. Passive air
sampling of organochlorine pesticides in Mexico. Environ Sci Technol 2009;43:
704–10.
Zar JH. Biostatistical analysis. 4th ed. Delhi, India: Pearson Education (Singapore) Pte Ltd;
2004.
Zhang L, Ma J, Venkatesh S, Li Y-F, Cheung P. Modeling evidence of episodic intercontinental long-range transport of lindane. Environ Sci Technol 2008;42:8791–7.