Click
Here
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112, D03103, doi:10.1029/2006JD007195, 2007
for
Full
Article
Atmospheric controls on the annual cycle of North African dust
Sebastian Engelstaedter1 and Richard Washington1
Received 13 February 2006; revised 10 July 2006; accepted 10 August 2006; published 13 February 2007.
[1] Dust emitted from desert regions and transported in the atmosphere has been
recognized for its potential to alter the Earth’s climate and environments. Satellite data
show that the largest source regions of dust (i.e., hot spots) are located in dry,
nonvegetated areas of West Africa and central Chad. Dust emissions from these sources
follow a distinct seasonal cycle. Whereas our understanding of processes controlling the
dust cycle of the Chad dust source has been improved through recent studies, our
understanding of the West African sources is limited because of the remoteness of the
sources and lack of surface observations. Using a satellite-derived dust index and
reanalysis atmospheric fields, we show that the annual dust cycle at the West African
dust hot spots is not related to changes in mean surface wind strength but is linked to
small-scale high-wind events. We find that the annual dust cycle correlates well with
changes in near-surface convergence associated with the annual north-south movement
of the Inter-Tropical Convergence Zone (ITCZ). Dust emissions in West Africa are highest
in June coinciding with the crossing of the convergence zone on its northward bound
over the dust hot spots. The increase in convergence leads to enhanced surface gustiness
suggesting that dry convection associated with an increase in the occurrence of small-scale
high-wind events and vertical velocity are the main processes controlling the annual
dust cycle at the West African dust sources.
Citation: Engelstaedter, S., and R. Washington (2007), Atmospheric controls on the annual cycle of North African dust, J. Geophys.
Res., 112, D03103, doi:10.1029/2006JD007195.
1. Introduction
[2] Mineral aerosols from the world’s desert regions have
the potential to impact the climate system by altering the
Earth’s radiation budget [Alpert et al., 2000; Andreae et al.,
2005; Kaufman et al., 2002] by scattering and absorbing
incoming solar radiation, absorbing terrestrial radiation and
changing the physical properties of clouds, such as brightness, spatial extent, lifetime or their ability to produce
rainfall [Levin et al., 1996; Rosenfeld et al., 2001; Wurzler
et al., 2000]. The magnitude of the radiative forcing and
even the sign (whether the dust forcing will result in a
cooling or warming of the Earth’s surface) are highly
uncertain with estimates ranging from 0.6 to +0.4 Wm 2
[Ramaswamy et al., 2001]. Possible feedback mechanisms
can increase or decrease the radiative forcing effect, although some recent studies suggest the dust has a net
cooling effect [Myhre et al., 2003; Perlwitz et al., 2001].
[3] An estimated half of the world’s atmospheric dust
originates from the North African deserts with emission
estimates ranging from 160– 760 106 tonnes per year
[Goudie and Middleton, 2001, and references therein], with
a recent study suggesting 1600 106 tonnes per year [Ozer,
2001]. Because dust can travel long distances in the
1
Climate Research Laboratory, Oxford University Centre for the
Environment, Oxford, UK.
Copyright 2007 by the American Geophysical Union.
0148-0227/07/2006JD007195$09.00
atmosphere, it can affect natural and human environments
far away from its source. Dust from North Africa is thought
to impact ecosystems off the American coast [Shinn et al.,
2000; Walsh and Steidinger, 2001] and in Puerto Rico
[Stallard, 2001], may alter biogeochemical cycles in the
Amazon Basin [Swap et al., 1992] and the Atlantic Ocean
[Jickells et al., 2005; Talbot et al., 1986], and may be a
potential threat to human health in the Sahel [Sultan et al.,
2005] and North America [Prospero, 1999; Shinn et al.,
2003]. The Saharan Air Layer (SAL), an elevated layer of
air and mineral Saharan dust between about 500–850 hPa,
may also impact the development of tropical cyclones which
occur over the tropical Atlantic between 10°N and 20°N by
inhibiting their ability to intensify [Dunion and Velden, 2004].
[4] Because of the important role that dust might play in
future climate change and its potential to impact the Earth’s
ecosystems and natural and human environments, it is
important to know where the major dust sources are, how
dust concentrations vary in space and time and what
controls this variability. Because of advancements in satellite technology, we now have a better understanding of
where major dust sources in North Africa are and where the
dust is transported within the atmosphere [e.g., Brooks and
Legrand, 2000; Chiapello and Moulin, 2002; Israelevich et
al., 2002; Middleton and Goudie, 2001; Moulin et al., 1998;
Prospero et al., 2002; Washington et al., 2003]. Major dust
producing areas in North Africa, as identified by the longterm mean of the satellite derived Total Ozone Mapping
Spectrometer absorbing Aerosol Index (TOMS AI)
D03103
1 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 1. Long-term mean TOMS AI (10) over Africa north of the equator (filled contours) calculated
using data from 1980 – 1992. Long-term mean (1961 –1990) precipitation (black isohyets) is derived from
a 0.5° by 0.5° rainfall data set [New et al., 1999]. West African major dust hot spots are indicated as WA1,
WA2, and WA3 and the Bodélé Depression in Chad as BOD.
[Herman et al., 1997], are a region in central Chad known
as the Bodélé Depression (marked as BOD) and a large
region in West Africa covering areas of Mauritania, Mali
and Algeria (three hot spots marked as WA1, WA2 and
WA3) (Figure 1).
[5] Both satellite and surface observations show that
North African dust emissions follow a distinct seasonal
cycle with periods of high dust activity (e.g., western Africa
in early summer) and periods that appear almost dust free
(e.g., November– December) (Figure 2). Surface observations from meteorological stations, such as visibility (in
desert regions dependent on atmospheric dust) and dust
event frequency, have also been used to describe the annual
dust cycle [Bertrand et al., 1979; Goudie and Middleton,
2001; Hoidale et al., 1977; Littmann, 1991; Middleton,
1986; N’tchayi Mbourou et al., 1997; Ozer, 2001].
However, most meteorological stations are located in the
Sahel or on the West African coast and not near the major
dust sources in the central Sahara. The controls on the
annual dust cycle of the West African sources are not well
understood although synoptic systems such as African
Easterly Waves, which are most prominent from June to
October, have been associated with dust production in
summer [Jones et al., 2003, 2004]. To fully explain the
observed seasonal dust variability in space and time, however, remains difficult as primary processes and drivers in
these regions have not yet been identified. As the most
intense dust sources are located in the hyperarid regions of
the Sahara where rainfall is very low (<200 mm yr 1,
Figure 1), it is likely that the annual dust cycle in these
regions is controlled by changes in near-surface winds. In
this paper we pose the following research question with a
focus on the West African dust sources: Is the surface wind
strength the primary control of the annual dust cycle? If not,
what other circulation parameters relate more closely to the
annual dust cycle?
2. Data and Methods
[6] As a measure of amounts of dust in the atmosphere
we use the absorbing Aerosol Index (AI) derived from the
Total Ozone Mapping Spectrometer (TOMS) which flew
on the Nimbus 7 satellite taking measurements from
November 1978 to May 1993 [Herman et al., 1997]. The
instrument allows the detection of absorbing aerosols
through the spectral difference between the 340 nm and
380 nm UV channels. An advantage of the TOMS AI product
is that it provides information about dust distributions over
land and ocean surfaces over a reasonably long period
(continuous 13 years record from the Nimbus 7 satellite).
[7] Despite the successful application of the TOMS AI in
the research of many aspects of dust in the climate system
[e.g., Alpert et al., 2000; Barkan et al., 2004; Borbely-Kiss
et al., 2004; Chiapello and Moulin, 2002; Israelevich et al.,
2002; Prospero et al., 2002; Washington et al., 2003] the
data set has some limitations. There is an ongoing debate
about how well the TOMS AI is able to capture dust near
the surface below about 1.5 km. Herman and Celarier
[1997] argue that the method applied to derive the TOMS
AI is not able to detect low-level UV-absorbing aerosols
near the surface. Torres et al. [2002], however, argue that
‘colored’ aerosols such as mineral dust can be detected
2 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 2. Mean monthly TOMS AI (10) calculated using data from 1980 – 1992.
close to the surface. The potential height bias would
underestimate atmospheric dust content in regions with
low-level dust transport. The altitude of dust aerosols has
an effect on the backscattered radiation, which affects the
method used to calculate the TOMS AI (difference between
340 nm and 380 nm UV channels) [see Torres et al., 1998].
Assuming that dust aerosols mix within the boundary layer,
the TOMS AI is therefore sensitive to the boundary layer
height (BLH) [Mahowald and Dufresne, 2004; Torres et al.,
1998]. Mahowald and Dufresne [2004] suggest that the
TOMS AI’s sensitivity to the BLH leads to an underestimation of source regions on the edges of deserts (e.g.,
Sahel) where the BLH is generally lower than in the center
of the desert. The TOMS AI is a column integrated measure
of the atmospheric dust content and therefore provides
information about the horizontal distribution of dust. No
information can be inferred from TOMS AI about the
vertical distribution of dust in the atmosphere. Because of
the overpass cycle of the satellite, measurements are available only once a day at local noon. The TOMS AI product is
available globally on a 1.25° longitude by 1° latitude grid.
For the analyses in this paper we use daily records of the
TOMS AI version 8 from January 1980 to December 1992
available at http://toms.gsfc.nasa.gov.
[8] Data on near-surface atmospheric conditions (10 m
surface wind speed, 10 m surface wind gusts and nearsurface divergence) are derived from European Centre for
Medium-Range Weather Forecasts (ECMWF) 40 year
reanalysis (ERA-40). Surface wind speeds are computed
using u and v components. Wind speed and wind gust data
are based on ERA-40 Surface/Single Levels. The ERA-40
divergence field consists of 60 vertical levels (sigma levels).
To represent near-surface divergence we use data from the
lowest (60th) level of the divergence field which, for
example, refers to 1012 hPa when the surface pressure is
1015 hPa. All parameters were interpolated to 1.25° 1°
grid resolution (resolution of TOMS AI data). Despite ERA40 data being available in 6-hourly time steps (00, 06, 12,
and 18 hours), we only use data representing 12 UTC to
match the measurements of the Nimbus 7 satellite as closely
as possible. At the prime meridian TOMS AI and ERA-40
data match best as both represent local noon. With increasing distance from the prime meridian west and east, the
difference between local noon (TOMS AI) and 12 hours
UTC (ERA-40) increases. However, over North Africa
these differences are likely to be negligibly small as the
maximum time difference is just about 3 hours (toward the
east). Surface wind gusts are available as 6 hour forecast
time steps which means that, in contrast to reanalysis data
which are available in 6-hourly time steps, a field for
12 hours is represented by the 6 hour forecast from 6 hours.
[9] Standard statistical methods were applied in this
study. As we do not necessarily expect linear relationships
between the annual cycles of dust and near-surface atmospheric conditions, correlation coefficients were calculated
using Spearman rank correlation (rs) [Wilks, 1995]. Correlation coefficients (rs) between the mean annual cycle of
daily noon TOMS AI and daily noon values of surface wind
speed, near-surface divergence and surface gustiness are
calculated by computing two time series each consisting of
365 values corresponding to each day of the year. Every day
is averaged over 13 years (1980 – 1992). A simple harmonic
regression model is used following Piegorsch and Bailer
[2005] to analyze temporal and spatial characteristics of the
3 of 14
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
D03103
Figure 3. The phase of the annual cycle of the TOMS AI. White areas indicate that the first harmonic is
not dominating. The locations of the major dust source are marked.
annual cycle of dust over North Africa. The application of
the model enables us to calculate the first harmonic phase of
the annual dust cycle, a measure of the time of peak dust
emissions.
3. Annual Dust Cycle
[10] North African dust emission follows a distinct seasonal cycle (Figure 2). In winter (November – February)
dust is transported in the northeasterly trade winds from
the Saharan desert toward the Gulf of Guinea where the dust
mixes with black carbon aerosols from biomass burning
resulting in a relatively strong TOMS AI signal near the
Guinea coast [Herman et al., 1997]. The most persistent
dust hot spot is the Bodélé Depression in central Chad,
which is active throughout most of the year and is one of the
most intense and persistent dust sources in the world
[Barkan et al., 2004; Prospero et al., 2002; Washington et
al., 2003; Washington et al., 2006]. Although in the TOMS
AI data the Bodélé Depression peaks in May (Figure 2),
Washington and Todd [2005] show that, despite some
differences between the data sets (TOMS AOT, TOMS AI
and dust plume frequency), dust emissions in the Bodélé
Depression are strongest at the beginning of the year
(January – March) with a secondary peak in October. The
offset in the TOMS AI may be explained by the potential
height bias described in Section 2. The major control of the
annul dust cycle in the Bodélé Depression is a northeasterly
low-level jet (LLJ), which is accelerated by the regional
orography. By sampling the two most dusty and least dusty
January, February, March, and April months between
1979 – 1992 (as represented by TOMS AOT) Washington
and Todd [2005] show that in the NCEP reanalysis data the
LLJ strengthens by about 3 ms 1 in high-dust months
relative to low-dust months.
[11] Apart from the Bodélé Depression, significantly less
dust production occurs during winter (January – March) in
the TOMS AI data (Figure 2). However, low-level dust
transport in winter from the Saharan desert toward the
northeastern Tropical Atlantic Ocean is evident from satellite
imagery (e.g., MODIS) and surface observations of atmospheric dust concentrations at Sal Island (Cape Verde)
[Chiapello et al., 1995]. That this is not generally evident
in the TOMS AI may be due to the TOMS AI height bias.
Toward the summer, dust production in the Western Sahara
becomes very active with a peak in intensity in June and July,
covering large parts of central Mali, Mauritania and southern
Algeria. A local hot spot in eastern Sudan peaks in June.
Toward the end of the year dust activity decreases in all parts
of North Africa with November representing the time of year
with the lowest dust emissions. Some dust activity occurs in
the summer in central and northern Algeria (June to August)
and in central Libya, Egypt and Sudan (April to August).
[12] The monthly progression of dust loading through the
annual cycle is qualitatively evident from Figure 2. Next we
quantify the phase of the first harmonic of the annual cycle
(Figure 3). Two distinct features are evident in the phase of
the TOMS AI cycle over North Africa. First, south of
15°N, the phase in aerosol loadings has a south to north
gradient, with early February as the peak month at 7°N
and June as the peak month at 15°N. This south-north
gradient closely matches the seasonal advance of the tropical convection and rainfall (Figure 4) which is associated
with the movement of the Inter-Tropical Convergence Zone
(ITCZ). The seasonality of rainfall in the Sahel also controls
the occurrence of black carbon from biomass burning which
is known to influence the TOMS signal in the dry season
4 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
strengthening of the north easterly) in the transients propagating across North Africa and the Mediterranean.
4. Controls on Annual Cycle of Dust
Figure 4. The annual cycle of TOMS AI (10) (filled
contours) and rainfall (black isohyets) in mm yr 1 displayed
as a function of latitude from 5.5°N to 19.5°N taken at
6.5°W representing a section through the Sahel from the
southern Ivory Coast through Mali to the eastern border of
Mauritania. TOMS AI values were calculated over 1982 –
1990. Long-term mean (1961– 1990) precipitation (black
isohyets) is derived form a 0.5° by 0.5° rainfall data set
[New et al., 1999]. The dashed horizontal line marks the
latitudinal position of dust source WA1.
(November– February). Second, north of 15°N an eastwest orientation of the phase is evident (Figure 3). The
phase maximum progresses from a centre over southeast
Libya and Egypt near 25°E in May to reach the west coast
of Africa between the summer months of early June and late
July. Additionally, the phase maximum shifts eastward with
the seasons from 25°E (maximum in early June) to the
phase maximum of late July along the east coast of Sudan.
A similar seasonal pattern has also been reported by Moulin
et al. [1998] using Meteosat satellite based retrievals which
show that high values of dust optical thickness shift from
the eastern to the western Mediterranean Basin in the
summer (March – August). The change from N – S to E– W
direction in the phasing may point to two different control
mechanisms. The N– S directed phase change is clearly
controlled by precipitation associated with the ITCZ whereas
the E– W directed change is more complex. In part it may
relate to the preferred location of ridging (and hence
4.1. Sahel Rainfall
[13] In the Sahel, the movement of the ITCZ and associated rainfall pattern are likely related to the magnitude of
dust emissions. Emissions are highest in winter when the
Sahel is driest (although atmospheric dust mixes with
aerosols from biomass burning in this season). From January onward, the ITCZ starts moving north, bringing the
monsoon rain to the Sahel and thereby scavenging the
atmosphere of aerosols. The advancing rain from the south
leads to an increase in soil moisture and vegetation cover in
the southern parts of the Sahel thereby reducing dust
emission in the Sahel. This results in the development of
a distinct southern boundary of aerosols at about 7°N in
May which moves north and reaches its most northern
position at about 13°N in August before moving south in
September (Figure 2). The region north of the monsoon
rains, notably central western Africa, becomes active from
spring into the early summer peaking in June/July before
dust occurrence subsides until the end of the year.
[14] The relationship between dust and rainfall in the
Sahel is also illustrated in Figure 4 where the annual cycle
of rainfall is superimposed on the TOMS AI for a cross
section through the Sahel from the dust hot spot WA1 (at
18.5°N and 6.4°W) south to the coast of Ivory Coast (at
5.5°N and 6.4°W). In the south between 5 – 12°N black
carbon aerosols from biomass burning mix with dust from
the Sahara resulting in higher TOMS AI values mainly in
December to January during the driest time of the year.
Intensive rainfall in summer (up to 3000 mm yr 1 in
August/September) coincides well with low TOMS AI
values. Toward the north, the length of the rainy season
gets gradually shorter and rain intensity decreases, thereby
losing its suppressive influence on dust emissions. African
Easterly Waves (AEW) have been associated with the
generation and transport of dust in North Africa [Jones et
al., 2003, 2004]. The occurrence of AEWs peaks in boreal
summer when atmospheric dust in the Sahel is at a minimum and emissions in central Western Sahara are at a
maximum. This north-south gradient in dust occurrence
might be explained by mesoscale convective systems to
the south of the AEW which are associated with precipitation and the reduction in atmospheric dust whereas the
northern part of the AEW is almost precipitation free,
resulting in no scavenging of atmospheric dust.
4.2. Near-Surface Atmospheric Circulation in
Central Sahara
4.2.1. Near-Surface Wind Speed
[15] All major dust sources (WA1, WA2, WA3 and BOD)
are located between 15–27°N in the hyperarid regions of
the Saharan desert where there is little or no vegetation
(Figure 1). Given the assumption that erodibility factors that
can limit dust emissions (such as the availability of deflatable sediments or vegetations cover) do not change significantly in these regions over the course of the year, dust
production is, to a large degree, a function of the surface
wind speed. From this perspective, we expect a positive
5 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 5. Correlation rs between the mean annual cycle of daily noon TOMS AI and daily noon ERA40 10 m surface wind speed. White areas indicate statistically insignificant correlations at the 0.05 level.
The locations of the major dust sources are marked. (In the black/white version, negative values are
indicated by dotted line contours and positive values are indicated by solid line contours.)
correlation between dust and wind speed over the annual
cycle so that an increase in surface wind speed enhances
dust production. To test this hypothesis, we have calculated
the correlation coefficient (rs) between the mean annual
cycle of daily noon TOMS AI and mean annual cycle of
daily noon surface wind speed as described in section 2 for
each grid box (Figure 5). The expected positive correlation
between the annual dust cycle and surface wind speed is
evident in large areas of Egypt, Libya, Algeria and near the
northwestern African coast as well as in a narrow band over
the Sahel between about 15 – 17°N (Figure 5). At the
location of the major dust sources, however, the relationship
is insignificant (e.g., WA3 or BOD). Only the sources WA1
and WA2 show a weak positive correlation. The annual
cycle of surface wind speed seems to match the annual cycle
of dust but not in the major dust hot spots. In the Sahel the
TOMS AI is affected by biomass burning aerosols in the dry
season, and therefore the correlation coefficients may not
accurately represent the relationship between dust and
surface wind speed in this region. It should be noted at this
point that the ERA-40 reanalysis data are initialized by
observations which, in the mostly uninhabited Saharan
desert, are only extremely sparsely distributed.
4.2.2. Near-Surface Divergence
[16] In trying to explain the relationship between surface
wind speed and dust at the dust hot spots, we examine the
seasonal changes in the general circulation over North
Africa. In January, northeasterly winds prevail over most
parts of North Africa. Convergence with southerly winds
from the equator occurs in the ITCZ, in a belt of low wind
speed, at about 7°N (Figure 6, top). As the summer
advances, the convergence belt moves north until August,
when it reaches its most northerly position between 16–
22°N (Figure 6, bottom). It then retreats south until
December. All the major dust sources are affected by the
crossing of the convergence belt at some time in the annual
cycle which raises the question of whether there is a
connection between the north-south movement of the convergence maxima and dust emissions.
[17] The correlation map of the mean annual cycle of
daily noon TOMS AI and the mean annual cycle of the daily
noon ERA-40 near-surface divergence (1980– 1992) (calculated as described in section 2) reveals that large areas of
North Africa show a negative correlation, so that low values
of divergence (corresponding to absolute convergence) are
associated with high values of TOMS AI (Figure 7). It is
plausible that in regions with strong surface convergence,
associated dry convection and increased vertical wind
velocity create conditions that favor dust emission and
transport into higher altitudes.
[18] Next we analyze in more detail the timing of peak
dust emissions at the West African dust hot spots in summer
in relation to the spatial movement of the near-surface
convergence. For each of the three West African dust
sources (WA1, WA2 and WA3) a latitudinal section
of the TOMS AI was taken between 10 – 30°N at the
corresponding longitudes (6.5°W, 4.5°W and 0.5°W, respectively) and contoured against time on the x-axis (Figure 8).
[19] At all three West African source regions (WA1, WA2
and WA3) the TOMS AI follows a clear seasonal cycle with
a peak in emissions in June (Figure 8). This peak in
emissions coincides with the crossing of the convergence
6 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
Figure 6. Monthly mean ERA-40 10 m surface wind vectors (1980 –1992) for January (top) and
August (bottom).
7 of 14
D03103
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 7. Correlation rs between the mean annual cycle of daily noon TOMS AI and the mean annual
cycle of daily noon ERA-40 near-surface convergence. White areas indicate statistically insignificant
correlations at the 0.05 level. The locations of the major dust sources are marked. (In the black/white
version, negative values are indicated by dotted line contours and positive values are indicated by solid
line contours.)
belt (Figure 8, solid line in all three panels) on its northward
movement over the dust source (Figure 8, dashed line in all
three panels). Note that despite the change of latitude from
18.5°N at WA1 to 21.5°N at WA3, this pattern is evident at
all three dust sources. This temporal and spatial coincidence
suggests that dry deep convection associated with surface
convergence plays an important role in dust generation in
the West African source regions. However, peak dust
emissions are restricted only to the crossing of the convergence belt on its northward advance and not when the
convergence zone retreats south about 2– 3 months later.
Where rainfall is sufficient to support vegetation, the lack of
emissions during the southward advance of the convergence
zone can be explained by a reduction in available sediments
as a result of increased vegetation cover. Reasons for the
lack of symmetry in the hyperarid Sahara are less clear, but
may relate to the stronger winds associated with higher
spring baroclinicity and the transient eddies which drive the
prevailing winds.
4.2.3. Near-Surface Gustiness
[20] The analysis of surface winds presented earlier was
based on noon wind values. The best associations between
dust and circulation in the West African dust hot spots are
with the passage of near-surface convergence. Turbulence is
likely to be an important characteristic of a boundary layer
featuring strong near-surface convergence, and turbulence
may relate more closely to dust generation than an instantaneous value of wind. The correlation map of the mean
annual cycle of daily noon TOMS AI and the mean annual
cycle of daily noon ERA-40 10 m surface wind gusts
(calculated as described in section 2) reveals that, with the
exception of some regions at the West African coast, in
central Chad and central Sudan, the whole Sahara shows a
significant positive correlation ranging at the West African
dust hot spots form 0.25 (WA3) to 0.625 (WA1 and WA2),
suggesting that near-surface turbulence is an important
driver of dust emissions (Figure 9).
4.2.4. Composite View of Dust Hot Spots
[21] In this section we consider the relationship between
dust and the parameters describing the near-surface circulation together at the three West African dust hot spots
(Figure 10). The annual cycle of TOMS AI is very similar at
all three dust hot spots with lowest values between 0.1 –1.5
in winter (January/December) and highest values between
3.5– 4.0 in summer (June/July). The annual cycle of wind
speed at WA1 and WA2 shows high values in winter
between 4.5– 5.5 ms 1 when TOMS AI values are lowest.
Apart from a coinciding decrease in surface wind speed and
TOMS AI between July and September at WA1 and WA2,
the annual cycle of wind speed does not match the annual
cycle of dust. At WA3 there is a strong mismatch between
the annual cycle of wind speed and TOMS AI (Figure 10).
The annual cycle of gustiness, however, matches the annual
cycle of TOMS AI well at the sources WA1 and WA2. At
the source WA3 the relationship is much weaker. At WA1
and WA2 gustiness peaks in the same month (June) as
divergence suggesting that convergence and gustiness are
linked. At the time of the second peak of convergence
(September at WA1 and WA2), which corresponds to the
second crossing of the convergence belt over the dust
8 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 8. TOMS AI (10) (filled contours) and latitudinal position of maximum near-surface
convergence (solid line superimposed on filled contours), for Saharan dust hot spots WA1 (top), WA2
(middle), and WA3 (bottom). The latitudinal positions of convergence (solid line superimposed on filled
contours) are defined by the maximum ratio between monthly mean ERA-40 10 m surface wind speed
and monthly mean noon 10 m surface gustiness between 10– 25°N. The latitudinal position of
convergence is calculated using the average between 1.5° and +1.5° of the respective longitudinal
position of the dust source. During winter (November– March) the latitudinal position of convergence
could not be calculated as at this time of year the position is not very well defined. The horizontal dashed
lines superimposed on the TOMS AI contours represent the latitudinal position of the dust sources.
source, gustiness is relatively low and wind speed is at its
annual minimum. Again, why the crossing of convergence
during its southward movement results in such a different
behavior of surface wind speed, gustiness, and dust emissions is not clear. Convergence at WA3 does not show a
double peak but matches the annual dust cycle well. WA3 is
located further north (21.4°N) than WA1 and WA2 (18.5°N
and 19.5°N respectively). The two crossings of convergence
occur in a shorter time and have probably been smoothed by
producing the long-term average. Surface wind speed and
gustiness at WA3 show a different pattern compared to WA1
and WA2.
5. Discussion and Summary
[22] In this paper we have analyzed the relationship
between the annual cycle of dust and the near-surface
circulation for the major North African dust sources. We
have shown that the annual cycle of dust does not match the
annual cycle of surface wind speed but that the agreement
with the annual cycle of near-surface convergence and
gustiness is strong. Both wind speed and gustiness are
measures of wind. In the ERA-40 data set, wind speed
variations with scales larger than both the spatial resolution
and the time step are assumed to be resolved by the
atmospheric model and are represented by the parameter
wind speed. The parameter gustiness accounts for wind
variability at scales comparable to or lower than the model
resolution (ECMWF, Part VII: ECMWF Wave-model documentation, in IFS documentation CY28r1, available at
http://www.ecmwf.int/research/ifsdocs/CY28r1/index.html,
edited 2004). This difference allows us to draw conclusions
about the importance of small-scale wind variability for the
generation of dust in desert environments. The timing of
peak dust emissions at the West African dust sources
coincides with the crossing of near-surface convergence
on its northward bound in summer (June) suggesting that
dry convection associated with an increase in near-surface
turbulence and enhanced vertical wind velocity is the
dominant process controlling the annual cycle of dust
production in these key regions.
[23] Dust is lifted from the ground when surface wind
velocity exceeds a certain threshold wind speed. At the West
African dust sources, the annual cycle of surface wind gusts
is very different from the annual cycle of surface wind
speed but relates well to the TOMS AI (except at WA3)
whereas wind speed shows no correlation to the TOMS AI
(Figure 10). This suggests that gustiness is a more representative parameter for the frequency of surface winds
exceeding the threshold wind speed and wind intensity at
the sources than mean wind speed. In the ECMWF data,
surface gustiness is derived as a function of wind friction
velocity (which depends on wind speed, surface roughness,
and atmospheric stability) and the magnitude of friction
velocity variability. As surface roughness does not change
significantly over the year in the dry nonvegetated source
9 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 9. Correlation rs between the mean annual cycle of daily noon TOMS AI and the mean annual
cycle of daily noon ERA-40 10 m surface gustiness. White areas indicate statistically insignificant
correlations at the 0.05 level. The locations of the major dust sources are marked. (In the black/white
version, negative values are indicated by dotted line contours and positive values by solid line contours.)
regions of West Africa, gustiness is likely to be a function of
friction velocity variability which seems to increase during
summer when convergence reaches the dust sources. In the
absence of fine resolution (space and time) wind data, it is
difficult to interpret precisely what features result in
enhanced gustiness over the key source regions. The agreement we see between gustiness and atmospheric dust
suggests that such observationally based studies are worth
doing.
[24] In June the maximum in surface solar heating moves
from the Sahel to the central Sahara establishing dry
convection whereas ageostrophic flow from the equator
brings moist humid air to the Sahel feeding the wet
convection in the ITCZ associated with the monsoon
rainfall [Cook, 1997, 1999; Parker et al., 2005; Sultan
and Janicot, 2000]. The resultant north-south gradient
between the relatively cooler temperatures to the south
and the hot Sahara to the north are also responsible for
the maintenance of the African Easterly Jet (AEJ) [Burpee,
1972; Cook, 1999; Parker et al., 2005; Thorncroft and
Blackburn, 1999]. The intense surface heating in summer is
the engine of the dry convection and convergence and
thereby indirectly influences dust emission at the West
African dust sources. Near-surface convection, if occurring
over surfaces that allow the emission of dust, leads to the
development of dusty convective plumes and dust devils,
phenomena frequently observed in desert regions [e.g., Idso,
1974; McTainsh and Pitblado, 1987; Renno et al., 2004].
The importance of convective processes for dust generation
has, however, only recently been recognized by Koch and
Renno [2005] who estimate that on a global scale, convec-
tive plumes and vortices contribute to about 35% of the
global dust budget. Convective processes due to intense
surface heating also play an important role in dust generation on Mars [Ferri et al., 2003; Renno et al., 2004].
[25] The data linking interannual dustiness and surface
convergence shows a reasonable though not perfect match
(Figure 8). Haywood et al. [2005] have shown that specific
dust outbreaks in the Sahara can lead to numerical model
longwave radiative errors of up to 50 Wm 2. It is reasonable to expect that the intensity and location of surface
heating and convergence could be strongly influenced by
aerosols. At least part of the reason for the inexact match
between the latitude of convergence and dust apparent in
some years might lie with the absence of mineral aerosols in
the ERA-40 data and hence the poor simulation of the
convergence field. From this perspective the agreement in
Figure 8 between dustiness and convergence is surprisingly
good.
[26] The intense surface heating in West Africa in summer is likely to result in an increase in surface turbulence
and gustiness as well as in enhanced convection and a
deepening of the planetary boundary layer. Mahowald and
Dufresne [2004] show that the TOMS AI is sensitive to the
height of the boundary layer (see section 2). From this
perspective one could argue that the good agreement
between peak dust emissions and the crossing of convergence over the dust sources (Figure 8) is not the result of an
increase in surface gustiness but the result of the TOMS
AI’s sensitivity to the boundary layer height and that the
gustiness just occurs contemporaneously with the increase
in the boundary layer depth. Figure 11 shows a latitudinal
10 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 10. Annual cycle of TOMS AI (solid line, all three panels) and ERA-40 10 m surface wind
speed (dashed line, top), ERA-40 10 m surface gustiness (dashed line, middle) and ERA-40 near-surface
divergence (dashed line, bottom). Monthly mean values for all data sets were calculated over the period
1980 –1992. Note that the right hand scale for divergence (bottom) is reversed.
time section of the Infrared Difference Dust Index (IDDI), a
dust aerosol product derived from Meteosat brightness
temperature measurements [Legrand et al., 2001],
taken between 15 – 25°N and averaged between 4 – 7°W
(corresponding to a region of peak emissions in IDDI which
covers the longitudes of WA1 and WA2). Although the
association is not as clear as that found with the TOMS AI
(Figure 8), the IDDI peaks in summer when the convergence (solid line superimposed on color contours) is located
close to the source (horizontal dashed line). The very similar
results derived from the use of a dust aerosol data set
independent from the TOMS AI and which does not have
a dependency on the boundary layer height, suggests that
the good agreement we see in Figure 8 is more likely to
result from convective turbulence and related surface gustiness than simply an artifact of boundary layer depth. It
remains to be seen whether this relationship holds when
other independent satellite derived measurements become
long enough to test it.
[27] Once dust is suspended in the atmosphere it is
transported into higher altitudes by an increased vertical
velocity due to dry convection. The evolution of vertical
velocity during the monsoon over West Africa shows a
significant increase in upward motion at around 20°N in
summer due to surface heating and convection [Fontaine et
al., 2002]. This way dust can be transported to higher
altitudes where it then is transported westward within a
fairly stable layer of air above the moist trade winds known
as the Saharan Air Layer (SAL) [Carlson and Prospero,
1972; Prospero and Carlson, 1972; Dunion and Velden,
2004]. The SAL forms in boreal summer over North Africa
by heated convective air (into which dust gets mixed) up to
6 km high which extends westward in a layer between 1.5
and 4.5 km (600– 800 hPa) characterized by an easterly
flow [Carlson and Prospero, 1972]. Petit et al. [2005]
11 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
D03103
Figure 11. IDDI (filled contours) and latitudinal position of maximum near-surface convergence (solid
line superimposed on filled contours) for the West Saharan dust source region located at about 5.5°W and
21.5°N. The latitudinal position of convergence (solid line superimposed on filled contours) is defined by
the maximum ratio between monthly mean ERA-40 10 m surface wind speed and monthly mean noon
10 m surface gustiness between 10 –25°N. The monthly IDDI and the latitudinal position of convergence
are calculated using the average between 4– 7°W. During winter (November– March) the latitudinal
position of convergence could not be calculated as at this time of year the position is not very well
defined. The horizontal dashed line superimposed on the TOMS AI contours represents the latitudinal
position of the dust source region.
show, using back trajectory analysis from Guadeloupe
(Caribbean), that over northern Mali (at about 3°W and
20°N, near dust source WA2) between 14– 15 June 1994
(note that June is the peak month in the West African dust
cycle, Figure 2 and Figure 10) dusty air masses were lifted
from about 600 m in the boundary layer to about 5000 m.
On the basis of the subsidence rates during the further
westward transport of the air parcel in the Saharan Air
Layer, they conclude that only high-altitude uplift of dust
occurring over Africa in summer has the potential to cross
the Atlantic and reach the Caribbean.
[28] Our results suggest that small-scale atmospheric processes that are likely to be associated with dry convection are
important for the emission of dust but they are usually
difficult to represent in numerical models. That small-scale
circulation processes near the surface have such an influence
on dust emission is also an important problem for dust
modelers. Offline dust models are used to simulate emission,
transport and deposition of dust. These models use prescribed
reanalysis winds with resolutions such as 1.125° 1.125°
(ECMWF) or 2° 2.5° (NCEP). These mean surface wind
fields are averaged over large areas (depending on resolution)
and do not account for subgrid-scale circulation such as dust
devils or convective dust plumes. Mean wind fields can
underestimate peak wind speeds at the surface during dust
events by up to 50% compared with observations [Tegen,
2003; Tegen and Miller, 1998]. Attempts have been made to
account for small-scale high-wind events by including a term
for probability distributions of wind speed variability [e.g.,
Grini et al., 2005]. Cakmur et al. [2004] parameterize
subgrid-scale winds in an atmospheric general circulation
model showing that subgrid wind variability, which is dominated by dry convection, significantly improves dust emissions when compared to satellite retrievals of dust. The good
agreement between the annual cycle of dust and ERA-40
surface gustiness in one of the largest source regions in the
world may give some further indication on how to improve
future dust model simulations.
[29] Acknowledgments. We would like to thank A. Stephens from
BADC for making ERA-40 data (especially surface wind gusts) available
and I. Tegen for useful comments.
References
Alpert, P., J. Herman, Y. J. Kaufman, and I. Carmona (2000), Response of
the climatic temperature to dust forcing, inferred from total ozone mapping spectrometer (TOMS) aerosol index and the NASA assimilation
model, Atmos. Res., 53(1 – 3), 3 – 14.
Andreae, M. O., C. D. Jones, and P. M. Cox (2005), Strong present-day
aerosol cooling implies a hot future, Nature, 435(7046), 1187 – 1190.
Barkan, J., H. Kutiel, and P. Alpert (2004), Climatology of dust sources in
North Africa and the Arabian Peninsula, based on TOMS data, Indoor
Built Environ., 13(6), 407 – 419.
Bertrand, J., A. Cerf, and J. L. Domergue (1979), Repartition in space and
time of dust haze south of the Sahara, WMO, 538, 409 – 415.
Borbely-Kiss, I., A. Z. Kiss, E. Koltay, G. Szabo, and L. Bozo (2004),
Saharan dust episodes in Hungarian aerosol: Elemental signatures and
transport trajectories, J. Aerosol Sci., 35(10), 1205 – 1224.
Brooks, N., and M. Legrand (2000), Dust variability over northern Africa
and rainfall in the Sahel, in Linking Land Surface Change to Climate
Change, edited by S. J. McLaren and D. Kniverton, pp. 1 – 25, Springer,
New York.
Burpee, R. W. (1972), Origin and structure of easterly waves in lower
troposphere of North Africa, J. Atmos. Sci., 29(1), 77 – 90.
Cakmur, R. V., R. L. Miller, and O. Torres (2004), Incorporating the effect
of small-scale circulations upon dust emission in an atmospheric general
circulation model, J. Geophys. Res., 109, D07201, doi:10.1029/
2003JD004067.
Carlson, T. N., and J. M. Prospero (1972), The large scale movement of
Saharan air outbreaks over the northern equatorial Atlantic, J. Appl.
Meteorol., 16, 1368 – 1371.
Chiapello, I., and C. Moulin (2002), TOMS and METEOSAT satellite
records of the variability of Saharan dust transport over the Atlantic
during the last two decades (1979 – 1997), Geophys. Res. Lett., 29(8),
1176, doi:10.1029/2001GL013767.
Chiapello, I., G. Bergametti, L. Gomes, B. Chatenet, F. Dulac, J. Pimenta,
and E. S. Suares (1995), An additional low layer transport of Sahelian
and Saharan dust over the north-eastern Tropical Atlantic, Geophys. Res.
Lett., 22(23), 3191 – 3194.
Cook, K. H. (1997), Large-scale atmospheric dynamics and Sahelian precipitation, J. Clim., 10(6), 1137 – 1152.
Cook, K. H. (1999), Generation of the African easterly jet and its role in
determining West African precipitation, J. Clim., 12(5), 1165 – 1184.
Dunion, J. P., and C. S. Velden (2004), The impact of the Saharan air layer
on Atlantic tropical cyclone activity, Bull. Am. Meteorol. Soc., 85(3),
353 – 365.
12 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
Ferri, F., P. H. Smith, M. Lemmon, and N. O. Renno (2003), Dust devils as
observed by Mars Pathfinder, J. Geophys. Res., 108(E12), 5133,
doi:10.1029/2000JE001421.
Fontaine, B., N. Philippon, S. Trzaska, and P. Roucou (2002), Spring to
summer changes in the West African monsoon through NCEP/NCAR
reanalyses (1968 – 1998), J. Geophys. Res., 107(D14), 4186,
doi:10.1029/2001JD000834.
Goudie, A. S., and N. J. Middleton (2001), Saharan dust storms: Nature and
consequences, Earth Sci. Rev., 56(1 – 4), 179 – 204.
Grini, A., G. Myhre, C. S. Zender, and I. S. A. Isaksen (2005), Model
simulations of dust sources and transport in the global atmosphere:
Effects of soil erodibility and wind speed variability, J. Geophys. Res.,
110, D02205, doi:10.1029/2004JD005037.
Haywood, J. M., R. P. Allan, I. Culverwell, T. Slingo, S. Milton,
J. Edwards, and N. Clerbaux (2005), Can desert dust explain the outgoing
longwave radiation anomaly over the Sahara during July 2003?, J. Geophys. Res., 110, D05105, doi:10.1029/2004JD005232.
Herman, J. R., and E. A. Celarier (1997), Earth surface reflectivity climatology at 340 – 380 nm from TOMS data, J. Geophys. Res., 102(D23),
28,003 – 28,012, doi:10.1029/97JD02074.
Herman, J. R., P. K. Bhartia, O. Torres, C. Hsu, C. Seftor, and E. Celarier
(1997), Global distribution of UV-absorbing aerosols from Nimbus 7/
TOMS data, J. Geophys. Res., 102(D14), 16,911 – 16,922, doi:10.1029/
96JD03680.
Hoidale, G. B., B. D. Hinds, and R. B. Gomez (1977), Worldwide data on
reduced visibility due to airborne dust, J. Opt. Soc. Am., 67(12), 1688 –
1690.
Idso, S. B. (1974), Tornado or dust devil: The enigma of desert whirlwinds,
Am. Sci., 62, 530 – 541.
Israelevich, P. L., Z. Levin, J. H. Joseph, and E. Ganor (2002), Desert
aerosol transport in the Mediterranean region as inferred from the TOMS
aerosol index, J. Geophys. Res., 107(D21), 4572, doi:10.1029/
2001JD002011.
Jickells, T. D., et al. (2005), Global iron connections between desert dust,
ocean biogeochemistry, and climate, Science, 308(5718), 67 – 71.
Jones, C., N. Mahowald, and C. Luo (2003), The role of easterly waves on
African desert dust transport, J. Clim., 16(22), 3617 – 3628.
Jones, C., N. Mahowald, and C. Luo (2004), Observational evidence of
African desert dust intensification of easterly waves, Geophys. Res. Lett.,
31, L17208, doi:10.1029/2004GL020107.
Kaufman, Y. J., D. Tanre, and O. Boucher (2002), A satellite view of
aerosols in the climate system, Nature, 419(6903), 215 – 223.
Koch, J., and N. O. Renno (2005), The role of convective plumes and
vortices on the global aerosol budget, Geophys. Res. Lett., 32, L18806,
doi:10.1029/2005GL023420.
Legrand, M., A. Plana-Fattori, and C. N’Doume (2001), Satellite detection
of dust using the IR imagery of Meteosat: 1. Infrared difference dust
index, J. Geophys. Res., 106(D16), 18,251 – 18,274, doi:10.1029/
2000JD900749.
Levin, Z., E. Ganor, and V. Gladstein (1996), The effects of desert particles
coated with sulfate on rain formation in the eastern Mediterranean,
J. Appl. Meteorol., 35(9), 1511 – 1523.
Littmann, T. (1991), Dust storm frequency in Asia – Climatic control and
variability, Int. J. Climatol., 11(4), 393 – 412.
Mahowald, N. M., and J.-L. Dufresne (2004), Sensitivity of TOMS aerosol
index to boundary layer height: Implications for detection of mineral
aerosol sources, Geophys. Res. Lett., 31, L03103, doi:10.1029/
2003GL018865.
McTainsh, G. H., and J. R. Pitblado (1987), Dust storms and related phenomena measured from meteorological records in Australia, Earth Surf.
Processes Landforms, 12(4), 415 – 424.
Middleton, N. J. (1986), The geography of dust storms, Ph. D. thesis, Univ.
of Oxford, Oxford, England.
Middleton, N. J., and A. S. Goudie (2001), Saharan dust: Sources and
trajectories, Trans. Inst. Br. Geogr., 26(2), 165 – 181.
Moulin, C., et al. (1998), Satellite climatology of African dust transport in
the Mediterranean atmosphere, J. Geophys. Res., 103(D11), 13,137 –
13,144, doi:10.1029/98JD00171.
Myhre, G., A. Grini, J. M. Haywood, F. Stordal, B. Chatenet, D. Tanré,
J. K. Sundet, and I. S. A. Isaksen (2003), Modeling the radiative impact
of mineral dust during the Saharan Dust Experiment (SHADE) campaign,
J. Geophys. Res., 108(D18), 8579, doi:10.1029/2002JD002566.
New, M., M. Hulme, and P. Jones (1999), Representing twentieth-century
space-time climate variability. Part I: Development of a 1961 – 90 mean
monthly terrestrial climatology, J. Clim., 12(3), 829 – 856.
N’tchayi Mbourou, G. N., J. J. Bertrand, and S. E. Nicholson (1997), The
diurnal and seasonal cycles of wind-borne dust over Africa north of the
equator, J. Appl. Meteorol., 36(7), 868 – 882.
Ozer, P. (2001), Les lithometeores en region sahelienne, Int. J. Trop. Ecol.
Geogr., 24, 1 – 317.
D03103
Parker, D. J., C. D. Thorncroft, R. R. Burton, and A. Diongue-Niang
(2005), Analysis of the African easterly jet using aircraft observations
from the JET2000 experiment, Q. J. R. Meteorol. Soc., 131(608), 1461 –
1482.
Perlwitz, J., I. Tegen, and R. L. Miller (2001), Interactive soil dust aerosol
model in the GISS GCM: 1. Sensitivity of the soil dust cycle to radiative
properties of soil dust aerosols, J. Geophys. Res., 106(D16), 18,167 –
18,192, doi:10.1029/2000JD900668.
Petit, R. H., M. Legrand, I. Jankowiak, J. Molinié, C. Asselin de Beauville,
G. Marion, and J. L. Mansot (2005), Transport of Saharan dust over the
Caribbean Islands: Study of an event, J. Geophys. Res., 110, D18S09,
doi:10.1029/2004JD004748.
Piegorsch, W. W., and A. J. Bailer (2005), Analyzing Environmental Data,
John Wiley, Hoboken, N. J.
Prospero, J. M. (1999), Long-range transport of mineral dust in the global
atmosphere: Impact of African dust on the environment of the southeastern United States, Proc. Natl. Acad. Sci. U. S. A., 96(7), 3396 – 3403.
Prospero, J. M., and T. N. Carlson (1972), Vertical and areal distribution of
Saharan dust over Western Equatorial North-Atlantic Ocean, J. Geophys.
Res., 77(27), 5255 – 5265.
Prospero, J. M., P. Ginoux, O. Torres, S. E. Nicholson, and T. E. Gill
(2002), Environmental characterization of global sources of atmospheric
soil dust identified with the NIMBUS 7 Total Ozone Mapping Spectrometer (TOMS) absorbing aerosol product, Rev. Geophys., 40(1), 1002,
doi:10.1029/2000RG000095.
Ramaswamy, V., O. Boucher, J. Haigh, D. Hauglustaine, J. Haywood,
G. Myhre, T. Nakajima, G. Y. Shi, and S. Solomon (2001), Radiative
forcing of climate change, in Climate Change 2001: The Scientific Basis,
edited by J. T. Houghton et al., pp. 349 – 416, Cambridge Univ. Press,
New York.
Renno, N. O., et al. (2004), MATADOR 2002: A pilot field experiment on
convective plumes and dust devils, J. Geophys. Res., 109, E07001,
doi:10.1029/2003JE002219.
Rosenfeld, D., Y. Rudich, and R. Lahav (2001), Desert dust suppressing
precipitation: A possible desertification feedback loop, Proc. Natl. Acad.
Sci. U. S. A., 98(11), 5975 – 5980.
Shinn, E. A., G. W. Smith, J. M. Prospero, P. Betzer, M. L. Hayes,
V. Garrison, and R. T. Barber (2000), African dust and the demise of
caribbean coral reefs, Geophys. Res. Lett., 27(19), 3029 – 3032,
doi:10.1029/2000GL011599.
Shinn, E. A., D. W. Griffin, and D. B. Seba (2003), Atmospheric transport
of mold spores in clouds of desert dust, Arch. Environ. Health, 58(8),
498 – 504.
Stallard, R. F. (2001), Possible environmental factors underlying amphibian
decline in eastern Puerto Rico: Analysis of US government data archives,
Conserv. Biol., 15(4), 943 – 953.
Sultan, B., and S. Janicot (2000), Abrupt shift of the ITCZ over West Africa
and intra-seasonal variability, Geophys. Res. Lett., 27(20), 3353 – 3356.
Sultan, B., K. Labadi, J. F. Guegan, and S. Janicot (2005), Climate drives
the meningitis epidemics onset in West Africa, PLoS Medicine, 2(1), 43 –
49.
Swap, R., M. Garstang, S. Greco, R. Talbot, and P. Kallberg (1992),
Saharan Dust in the Amazon Basin, Tellus, Ser. B, 44(2), 133 – 149.
Talbot, R. W., R. C. Harriss, E. V. Browell, G. L. Gregory, D. I. Sebacher,
and S. M. Beck (1986), Distribution and geochemistry of aerosols in the
tropical North Atlantic troposphere: Relationship to Saharan dust,
J. Geophys. Res., 91(D4), 5173 – 5182.
Tegen, I. (2003), Modeling the mineral dust aerosol cycle in the climate
system, Quart. Sci. Rev., 22(18 – 19), 1821 – 1834.
Tegen, I., and R. Miller (1998), A general circulation model study on the
inter-annual variability of soil dust aerosol, J. Geophys. Res., 103(D20),
25,975 – 25,995.
Thorncroft, C. D., and M. Blackburn (1999), Maintenance of the African
easterly jet, Q. J. R. Meteorol. Soc., 125, 763 – 786.
Torres, O., P. K. Bhartia, J. R. Herman, Z. Ahmad, and J. Gleason (1998),
Derivation of aerosol properties from satellite measurements of backscattered ultraviolet radiation: Theoretical basis, J. Geophys. Res., 103(D14),
17,099 – 17,110.
Torres, O., P. K. Bhartia, J. R. Herman, A. Sinyuk, P. Ginoux, and
B. Holben (2002), A long-term record of aerosol optical depth from
TOMS observations and comparison to AERONET measurements,
J. Atmos. Sci., 59(3), 398 – 413.
Walsh, J. J., and K. A. Steidinger (2001), Saharan dust and Florida red
tides: The cyanophyte connection, J. Geophys. Res., 106(C6), 11,597 –
11,612.
Washington, R., and M. C. Todd (2005), Atmospheric controls on mineral
dust emission from the Bodélé Depression, Chad: The role of the low
level jet, Geophys. Res. Lett., 32, L17701, doi:10.1029/2005GL023597.
Washington, R., M. Todd, N. J. Middleton, and A. S. Goudie (2003),
Dust-storm source areas determined by the total ozone monitoring spec-
13 of 14
D03103
ENGELSTAEDTER AND WASHINGTON: ANNUAL DUST CYCLE
trometer and surface observations, Ann. Assoc. Am. Geogr., 93(2), 297 –
313.
Washington, R., M. C. Todd, S. Engelstaedter, S. Mbainayel, and F. Mitchell
(2006), Dust and the low-level circulation over the Bodélé Depression,
Chad: Observations from BoDEx 2005, J. Geophys. Res., 111, D03201,
doi:10.1029/2005JD006502.
Wilks, D. S. (1995), Statistical Methods in the Atmospheric Sciences: An
Introduction, Elsevier, New York.
D03103
Wurzler, S., T. G. Reisin, and Z. Levin (2000), Modification of mineral dust
particles by cloud processing and subsequent effects on drop size distributions, J. Geophys. Res., 105(D4), 4501 – 4512.
S. Engelstaedter and R. Washington, Climate Research Lab, Oxford
University Centre for the Environment, South Parks Road, Oxford, OX1
3QY, UK. ([email protected]; richard.washington@
ouce.ox.ac.uk)
14 of 14