1 OCTOBER 2009
WONG ET AL.
5149
Maintenance of Lower Tropospheric Temperature Inversion in the Saharan
Air Layer by Dust and Dry Anomaly
SUN WONG AND ANDREW E. DESSLER
Department of Atmospheric Sciences, Texas A&M University, College Station, Texas
NATALIE M. MAHOWALD
Department of Earth and Atmospheric Sciences, Cornell University, Ithaca, New York
PING YANG AND QIAN FENG
Department of Atmospheric Sciences, Texas A&M University, College Station, Texas
(Manuscript received 19 September 2008, in final form 29 April 2009)
ABSTRACT
The role of Saharan dust and dry anomaly in maintaining the temperature inversion in the Saharan air layer
(SAL) is investigated. The dust aerosol optical thickness (AOT) in the SAL is inferred from the measurements taken by Aqua Moderate Resolution Imaging Spectroradiometer (MODIS), and the corresponding
temperature and specific humidity anomalies are identified using the National Centers for Environmental
Prediction (NCEP) data in August–September over the North Atlantic tropical cyclone (TC) main development region (MDR; 108–208N, 408–608W). The authors also study the SAL simulated in the National Center
of Atmospheric Research (NCAR) Community Atmosphere Model, version 3 (CAM3), coupled with dust
radiative effect. It is found that higher AOT is associated with warmer and dryer anomalies below 700 hPa,
which increases the atmospheric stability. The calculated instantaneous radiative heating anomalies from a
radiative transfer model indicate that both the dust and low humidity are essential to maintaining the temperature structure in the SAL against thermal relaxation. At 850 hPa, heating anomalies caused by both the dust
and dry anomalies (for AOT . 0.8) are 0.2–0.4 K day21. The dust heats the atmosphere below 600 hPa, while
the dry anomaly cools the atmosphere below 925 hPa, resulting in a peak of heating rate anomaly located at
700–850 hPa. In the eastern Atlantic, dust contributes about 50% of the heating rate anomaly. Westward of
408W, when the dust content becomes small (AOT , 0.6), the heating rates are more sensitive to the water
vapor profile used in the radiative transfer calculation. Retrieving or simulating correct water vapor profiles is
essential to the assessment of the SAL heating budgets in regions where the dust content in the SAL is small.
1. Introduction
The Saharan air layer (SAL) is a well-mixed layer of
dry, warm air located between the marine boundary
layer and the ;500-hPa level. From late spring to early
fall, the SAL is frequently advected westward with
Saharan dust across the North Atlantic Ocean (Carlson
and Prospero 1972; Dunion and Velden 2004; Karyampudi
et al. 1999; Nalli et al. 2005, 2006; Prospero and Carlson
1972; Wong and Dessler 2005; Wong et al. 2006). Using
Corresponding author address: Sun Wong, Department of Atmospheric Sciences, Texas A&M University, 3150 TAMU, College
Station, TX 77840-3150.
E-mail: [email protected]
DOI: 10.1175/2009JCLI2847.1
Ó 2009 American Meteorological Society
the Geostationary Operational Environmental Satellite
(GOES) to track the SAL across the North Atlantic
Ocean, Dunion and Velden (2004) found a phenomenological connection between the SAL and the suppression
of the tropical cyclone (TC) activity. They suggested that
the dissipation of deep convection occurring in the interior of the SAL plays a role in suppressing the development of TC activity. Using satellite-derived aerosol
optical thickness (AOT) and cloud brightness temperature, Wong and Dessler (2005) verified the suppression
of deep convections associated with the warm and dry
anomalies of the SAL. Because of the important role of
the SAL in TC activity and convection, it is necessary to
improve the current knowledge about the properties
and evolution of the SAL.
5150
JOURNAL OF CLIMATE
Wong et al. (2006, 2008) found that variations in summertime dust transport are associated with temperature
anomalies near 850 hPa. These temperature variations
drive the variations in the circulation at 700 hPa via the
thermal wind relation, which has previously been identified to be responsible for advecting Saharan dust off
the coast of West Africa (Carlson and Prospero 1972;
Colarco et al. 2003; Jones et al. 2004; Karyampudi and
Carlson 1988; Kaufman et al. 2005; Wong et al. 2006).
Because the dust and warm anomaly in the SAL are both
advected by the winds in the lower troposphere, it raises a
key question of what role, if any, the dust plays in maintaining the thermal structure of the SAL.
Although many studies have focused on dust radiative
forcing at the surface (srf) and the top of the atmosphere
(TOA; Carlson and Benjamin 1980; Evan et al. 2008;
Li et al. 2004; Myhre et al. 2003; Weaver et al. 2002;
Yoshioka et al. 2007; Yu et al. 2006), little effort has
been made to analyze the vertical distribution of heating
caused by dust. Carlson and Benjamin (1980) calculated
the dust radiative heating rates based on the observed
dust distribution over the eastern Atlantic and available
optical properties of dust. They noticed significant
heating rates by dust (.1 K day21) in the SAL, which
suggested that the dust plays an important role in stabilizing the atmosphere. Moreover, they found that the
maximum effect of dust on the heating rates is located
at ;700 hPa, where dust has the highest concentration.
However, the dust optical properties used in Carlson and
Benjamin (1980) are more absorptive than those found in
recent observations (Colarco et al. 2002; Dubovik et al.
2002; Kaufman et al. 2001; Sinyuk et al. 2003). Using the
Stratospheric Aerosol and Gas Experiment (SAGE) II
retrieved vertical dust profile, Zhu et al. (2007) estimated
the clear-sky net heating due to dust near the Saharan
coast to be about 0.2–0.3 K day21 below 3 km.
The intent of this study is to quantify the effects of
both the dust and the dry anomaly associated with the
SAL on the atmospheric heating rates, aimed at a better
understanding of their implications on the thermal
structure of the SAL. To do this, we first investigate the
thermal and moisture structure of the SAL. Then we
employ a radiative transfer model to estimate the instantaneous radiative heating anomaly caused by the
dust and dry anomaly in the SAL.
2. Data and models
In this study we use dust as a tracer to track the SAL
(Evan et al. 2006; Wong and Dessler 2005). The atmospheric temperature and specific humidity anomalies
associated with the SAL are obtained by binning the
reanalysis temperature profiles from the National Cen-
VOLUME 22
ter of Environmental Prediction (NCEP)/Department
of Energy Global Reanalysis 2 (NCEP-2; Kanamitusu
et al. 2002) according to the Moderate Resolution
Imaging Spectroradiometer (MODIS; Kaufman et al.
1997; King et al. 2003; Yu et al. 2004) AOT data. The
specific humidity from the NCEP final analyses (FNL),
which are forecast data and more consistent with the
dropsonde data of humidity and the GOES SAL tracking imagery (Dunion and Velden, 2004), is used in the
present radiative transfer calculation.
The MODIS aerosol retrieval algorithm is based on
six MODIS bands (0.55–2.1 mm), and the retrieved
aerosol properties are provided with a 10-km resolution
(Kaufman et al. 1997; Levy et al. 2003; Remer et al.
2002). Radiances at these wavelengths are essentially
due to reflected solar radiation and thus only available
during daytime. The AOT at 0.55 mm (hereafter AOT)
are archived in the MYD04 level-2 product (collection 5)
from the Aqua satellite, which has an overpass time of
approximately 1330 . We averaged the level-2 AOT data
into 2.58 3 2.58 boxes covering the tropical North Atlantic to match the horizontal grid of the NCEP datasets.
A sensitivity test has shown that the cloud fraction of the
chosen MODIS AOT pixels does not influence the results shown in this study. Hereafter, we pick those pixels
with a cloud fraction less than 75%.
MODIS AOT includes the contributions from many
types of aerosols. In the tropical and subtropical North
Atlantic Ocean, the main contribution to AOT is from
dust, maritime aerosol, and aerosol from biomass
burning in South Africa. MODIS retrievals do not distinguish the fractions of different types of aerosols. Using MODIS-measured fractions of fine aerosols in the
total AOT, Kaufman et al. (2005) determined the dust
contribution in the total MODIS AOT over the tropical
North Atlantic. Here we apply the same method to estimate the dust contribution of the MODIS-measured
AOT. The dust optical thickness (t du) is determined by
t du 5
[t( f an
f ) t ma ( f an
( f an f du )
f ma )]
,
(1)
where t and f are the MODIS AOT and fraction of
fine mode particles; fdu (50.5), fan (50.9), and fma (50.3)
are the empirical fractions of fine mode particles in dust,
anthropogenic, and marine aerosols, respectively; and
t ma is proportional to the NCEP surface wind speed
(50.007W 1 0.02, where W is the NCEP surface wind
speed). There may be uncertainties in the determined
dust AOT because of the uncertainties in the input parameters; however, because of the large amount of dust
(see Table 1) involved in our calculations, these uncertainties do not generate significant differences in our
1 OCTOBER 2009
5151
WONG ET AL.
TABLE 1. Averaged dust AOT contents in the various MODIS (at 0.55 mm) and CAM3 (at 0.64 mm) total AOT ranges over the TC MDR
for August–September in 2003–06. The numbers in the parentheses show the percentage content of dust in the total AOT.
AOT
Region I
MODIS
CAM3
Region II
MODIS
CAM3
Region III
MODIS
CAM3
0–0.3
0.3–0.4
0.4–0.5
0.5–0.6
0.6–0.8
.0.8
0.13 (65)
0.10 (43)
0.28 (80)
0.23 (66)
0.37 (82)
0.32 (71)
0.47 (85)
0.43 (78)
0.61 (90)
0.55 (83)
0.98 (93)
0.78 (90)
0.11 (61)
0.13 (57)
0.27 (77)
0.24 (69)
0.37 (82)
0.34 (76)
0.46 (84)
0.43 (80)
0.59 (87)
0.53 (83)
0.95 (91)
—
0.08 (53)
0.11 (58)
0.27 (77)
0.26 (76)
0.36 (80)
0.35 (80)
0.46 (84)
0.44 (83)
0.59 (87)
—
0.87 (90)
—
results. Hereafter, total AOT and dust AOT are referred to as AOT and dust AOT, respectively.
To estimate the radiative heating by dust, vertical
profiles of dust concentrations are necessary. We calculate the vertical distributions of dust in four size bins (0.1–
1.0, 1.0–2.5, 2.5–5.0, and 5.0–10 mm in diameters) from
the climatologically monthly mean of 1979–2000 from a
simulation of the Model of Atmospheric Transport and
Chemistry (MATCH; Rasch et al. 1997). The simulation
is driven by NCEP reanalysis meteorological fields (T62
resolution, ;1.88 3 1.88, and 28 vertical levels), which
have been extensively compared to observations for their
climatology distribution, as well as daily to interannual
variability (Luo et al. 2003; Mahowald et al. 2003, 2002;
Zender et al. 2003; Hand et al. 2004). The vertical distributions of the dust AOT at a visible wavelength
(0.64 mm) are used to estimate the vertical profiles of
radiative heating rate by dust. We rescale the relative
particle size distribution of dust to be larger so that it is
consistent with observations (Reid et al. 2003; Hand et al.
2004). We also lift the dust distribution by 100 hPa, so
that its peak falls in the climatological range of dust
layers (;2–4 km) observed in the Cloud–Aerosol Lidar
Infrared Pathfinder Satellite Observation (CALIPSO)
level-2 aerosol layer product (version 2.01; Winker et al.
2003) in August–September 2006.
Radiative heating rates are computed from a radiative
transfer model (hereafter Chou’s model) that was developed at the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center (Chou
and Suarez 2002; Chou et al. 2003). In this model,
shortwave reflection and transmission associated with a
dusty atmosphere are computed on the basis of the
d-Eddington approximation for 11 spectral bands (Chou
and Suarez 2002). The longwave radiation is largely accounted for in a parameterization form that includes the
absorption and scattering by aerosols for nine spectral
bands (Chou et al. 2003). The optical properties of
mineral dust (extinction coefficient, single-scattering albedo, and asymmetry factor) are specified as functions of
spectral bands and particle sizes in an approach similar to
those used by Mahowald et al. (2006) and Yoshioka et al.
(2007). The indices of refraction have been derived from
Patterson (1981) for visible wavelengths, Sokolik et al.
(1993) for the near-infrared spectrum, and Volz (1973)
for the infrared spectrum. The imaginary part of the
visible wavelength refraction indices have been linearly
interpolated from the observation-based estimates by
Sinyuk et al. (2003) and Dubovik et al. (2002) at 0.33,
0.36, 0.44, 0.66, 0.87, and 1.22 mm for the spectral region
of 0.33–0.97 mm. Table 2 shows the single-scattering albedos in the visible wavelength bands for the four size
bins. These values are consistent with those inferred from
satellite- and ground-based remote sensing data (Colarco
et al. 2002; Dubovik et al. 2002; Kaufman et al. 2001).
We also analyze simulations from the Community
Atmosphere Model, version 3 (CAM3; Collins et al.
2006), in which online dust transport and radiative feedbacks are implemented (Mahowald et al. 2006; Yoshioka
et al. 2007). These simulations have been compared to
available observations (Mahowald et al. 2006; Yoshioka
et al. 2007). The dust and the dry anomaly of the SAL in
the simulation are analyzed. We further perform offline
heating rate calculations using these simulated dust and
dry anomaly profiles. By comparing the dust, thermal,
and moisture simulations with those in the MATCH and
reanalysis data, we can evaluate the model’s ability for
simulating the SAL. Hereafter, we refer to the calculations that use the MODIS, MATCH, and NCEP datasets
TABLE 2. Single-scattering albedos of dust in the visible wavelength bands used in the radiative transfer model as a function of
dust size bin.
Spectral
range
(mm)
Dust
size
0.1–1 mm
Dust
size
1–2.5 mm
Dust
size
2.5–5 mm
Dust
size
5–10 mm
0.32–0.4
0.4–0.7
0.7–1.22
0.93
0.98
0.98
0.86
0.97
0.98
0.77
0.94
0.97
0.69
0.90
0.94
5152
JOURNAL OF CLIMATE
FIG. 1. August–September in 2003–06 climatology of total AOT
for (a) Aqua MODIS at 0.55 mm and (b) CAM3 at 0.64 mm. The
white area represents locations where no valid AOT retrievals
were made. The dotted lines indicate the TC MDR divided into
three subregions (I, II, and III).
as the ‘‘MODIS–MATCH’’ and those using CAM3
simulation as the ‘‘CAM3.’’
3. Results
a. Temperature and moisture anomalies in the SAL
Figure 1 shows the MODIS and CAM3 AOT climatologies for August–September from 2003 to 2006. Both
MODIS and CAM3 data indicate a large amount of
aerosols transported from the Saharan desert to the
tropical Atlantic. The geographical patterns of MODIS
and CAM3 AOTs are similar, although CAM3 simulates higher AOTs in the western Atlantic.
We divide the TC main development region (MDR;
108–208N, 208–608W) (Goldenberg et al. 2001) into three
subregions, as shown in Fig. 1. Region 1 covers the
longitude range of 208–308W, region 2 covers 308–408W,
and region 3 covers 408–608W. For each of these regions,
we associate each daily 2.58 3 2.58 MODIS AOT with
the collocated daily NCEP reanalysis temperature pro-
VOLUME 22
file. To provide the temperature response as a function
of AOT, the temperature profiles are then averaged for
various AOT bins. The optical thickness bins used are
0–0.3, 0.3–0.4, 0.4–0.5, 0.5–0.6, 0.6–0.8, and ,0.8. We
define the temperature anomaly profiles to be the difference of the averaged temperature profile in each bin
and the profile for AOT less than 0.3. The results are
shown in Figs. 2a–c. Similar analyses are applied to
CAM3’s AOT and temperature profiles, and the results
are shown in Figs. 2d–f.
In the NCEP reanalysis, there are warm anomalies in
all regions below ;600 hPa, with the maximum anomaly
located at 850 hPa in region 1 and rising to 700 hPa in
region 3 (Figs. 2a–c). Above the warm anomalies lie the
cold anomalies, with the maximum cold anomaly located at about 500 hPa. The anomalies simulated by
CAM3 are qualitatively similar but with a magnitude
about twice that found in NCEP reanalyses. Moreover,
the cold anomalies in the midtroposphere are missing in
the CAM3 simulation, and the maximum of the warm
anomaly in region 1 for AOT . 0.8 is located in a lower
altitude (925 hPa). The lifting of the warm anomaly
maximum across the Atlantic is also not evident in the
CAM3 simulation. In regions 2 and 3, CAM3 simulates
fewer large AOT events, and we omit any case with a
sampling size less than 10.
The SAL is also associated with dry air. To illustrate
this, we apply the same technique to compute the specific humidity anomalies as we did for the temperature,
and the results are shown in Fig. 3. All three regions
show dry anomalies below 300 hPa, reaching the maximum around 850–925 hPa in region 1 and 700 hPa in
region 3 for the NCEP data. The dry anomalies in the
CAM3 are about twice that found in the NCEP data (as
are the temperature anomalies in the lower troposphere
in the CAM3, see Fig. 2).
The inversion layers (the maximum temperature and
minimum humidity anomalies) shown in NCEP and
CAM3 data are both significant at the 99% confidence
level by Student’s t tests. The inversion layers rise in altitudes as they are transported from region 1 to region 3,
which is consistent with the rise of the SAL base when
it is transported westward (Karyampudi et al. 1999).
Dunion and Marron (2008) compared the rawinsonde
measurements of temperature and humidity for the nonSAL and SAL conditions in 2002 over the western
tropical Atlantic and the Caribbean Sea (west of 608W).
They found a temperature increase in 500–700 hPa and
maximum dry anomaly located at 500 hPa, consistent
with the rising SAL base shown in the NCEP reanalysis.
Dunion and Velden (2004) used GPS dropsondes to
estimate the dry anomaly in the SAL in the western
tropical Atlantic to be 2.5–5.5 g kg21 between 600 and
1 OCTOBER 2009
WONG ET AL.
5153
FIG. 2. (a)–(c) NCEP temperature profile anomalies (K) over the TC MDR as a function of
MODIS AOT ranges for August–September in 2003–06. (d)–(f) Same as (a)–(c) but for the
CAM3 temperature anomalies as functions of the CAM3 AOT. The anomalies are relative to
the temperature profiles with the smallest AOT range (0–0.3).
850 hPa and extend up to 300 hPa. Radiosonde observations of the SAL (Nalli et al. 2005; Nalli et al. 2006;
Szczodrak et al. 2007) also show similar dry anomaly
profiles in the SAL. Our dry anomalies estimated from
NCEP forecast data and the CAM3 simulation also extend up to about 300 hPa; however, our anomalies in
the lower troposphere over the western tropical Atlantic
are smaller in magnitude (NCEP’s 0.5–1 g kg21 and
CAM3’s 0.5–2.5 g kg21). This is because the coarse-grid
models average out spatial variation smaller than the grid
size and/or there may be biases in the model simulations.
b. Radiative heating flux anomaly due to dust
To understand the role that dust plays in the temperature anomaly in the SAL, we estimate the radiative
heating anomaly due to dust. Since the SAL contains
both dust and dry air, we also estimate the heating
anomaly caused by both the dust and the dry anomaly
(hereafter, the SAL heating anomaly). For all calculations, the background atmosphere is prescribed by the
averaged temperature and water vapor profiles for AOT
less than 0.3 in each region. The heating rate anomaly is
similarly defined as the difference in heating rate from
the heating rate in which AOT is less than 0.3. In this
study, we do not consider the effects of clouds on this
analysis, and all calculations are done for the clear-sky
conditions. Heating anomalies thus calculated are ‘‘instantaneous,’’ which represent the heating anomalies
instantaneously generated by the dust or dry anomaly
before any atmospheric response.
Since this study is to estimate the climatological heating effect of dust, we compute the daily averages of the
instantaneous heating rates. The daily averaged heating
and flux anomalies are the averages of these quantities
5154
JOURNAL OF CLIMATE
VOLUME 22
FIG. 3. Similar to Fig. 2 but for specific humidity profile anomalies (g kg21).
computed at 0000, 6000, 1200, and 1800 UTC by varying
the solar zenith angle and the corresponding temperature
and water vapor profiles. Since Aqua MODIS measures
AOT approximately near 1330 LT, we assume the diurnal AOT variation is small and use the same AOT for our
daily 4-time radiative transfer calculations.
We first consider the heating flux anomaly caused by
dust alone (keeping temperature and moisture profiles
at the background values). We applied Eq. (1) (Kaufman
et al. 2005) to estimate the dust contribution to the observed AOT for August–September in 2003–06. The
results are listed in Table 1. In the MDR, the dust contribution is more than 50% of the total AOT. Dust
contributes the smallest in the western tropical Atlantic
(i.e., region 3) for the total AOT range of 0–0.3, and the
largest (;92%) near the west coast of Africa (i.e., region
1) for the total AOT . 0.8. The CAM3’s dust content is
in agreement with MODIS’s, with the largest difference in dust AOT of about 0.2 located in 208–308W
when total AOT . 0.8. The CAM3’s dust in general
contributes smaller fraction to the total AOT, with the
largest difference in the dust percentage fraction to
the MODIS’s of about 22% located in 208–308W when
the total AOT , 0.3.
Figure 4 shows the vertical profiles of the models’ dust
AOT that are used in our radiative transfer calculations.
For this figure, the column-integrated AOT has been
scaled to one. The altitude range of dust layers measured
in CALIPSO is also indicated. Because we want to use
the MATCH profile as a baseline profile, we lifted the
MATCH dust AOT to match the CALIPSO dust layer
range located between 600 and 800 hPa (;2–4 km). This
altitude range is also in agreement with aircraft measurements of dust vertical profiles by Léon et al. (2003)
and Myhre et al. (2003). The CAM3 dust AOT peaks at
lower altitudes (800–900 hPa). In general, CAM3 simulates a dust layer with wider vertical extent than the
MATCH.
1 OCTOBER 2009
5155
WONG ET AL.
TABLE 3. Radiative forcing (W m22) by the MODIS–MATCH
dust at the surface and TOA over the TC MDR for August–
September in 2003–06. The column-integrated dust AOT at
0.55 mm for each subregion and the whole MDR is used for the
calculations. Forcing efficiency is the forcing divided by the corresponding regional-averaged dust AOT.
SWR
LWR
FIG. 4. Climatology of the (a) MATCH and (b) CAM3 dust
vertical profiles over three subregions in the TC MDR for August–
September in 2003–06. The dust concentrations are given as AOT
per 20 hPa. The dotted horizontal lines in (a) show the altitude
range of dust layers indicated by the CALIPSO version 2.01 layer
data over the MDR. The MATCH dust profile is lifted by 100 hPa
so that the dust concentration peak is within this altitude range.
The vertically integrated AOTs are scaled to one for both plots.
To estimate the effects of dust on the heating rate, we
put dust profiles into Chou’s radiative transfer model. The
input profile is based on the model’s simulated profile but
with column-integrated dust AOT at 0.64 mm (a visible
wavelength close to 0.55 mm at which MODIS reports its
AOT) scaled to be equal to the dust AOT values estimated from the MODIS retrievals (listed in Table 1). In
this way, we represent the vertical dust distributions for
the various total AOT ranges shown in Figs. 2, 3. Although the MATCH AOT is obtained at a slightly different wavelength from that of MODIS AOT, direct
measurements of dust AOT (e.g., Aerosol Robotic Network) at locations near the region of our interest (e.g.,
Cape Verde) indicate that the AOT at these two wavelengths differ by less than 10%. We also test that this
difference does not significantly influence our results
about the SAL heating property.
Dust AOT
TOA
TOA efficiency
Srf.
Srf. efficiency
TOA
TOA efficiency
Srf.
Srf. efficiency
Region
I
Region
II
Region
III
MDR
0.35
210.8
230.9
215.8
245.3
1.64
4.69
3.58
10.2
0.26
27.82
230.1
211.4
243.8
1.26
4.83
2.74
10.5
0.16
24.49
228.1
26.59
241.2
0.52
3.22
1.33
8.29
0.23
26.90
230.0
210.1
243.9
0.99
4.30
2.25
9.78
We first calculate dust radiative forcing at the surface
and TOA over the MDR. The forcings are defined as the
difference of net downward radiative fluxes for the
various regional-averaged column-integrated dust AOT
from the clear-sky flux. The results are summarized in
Tables 3 and 4, in which we have also computed the
forcing efficiency for dust aerosols, defined as the forcing divided by the column-integrated dust AOT. We see
that the forcing efficiency is approximately region independent (in particular at the TOA), providing a
means for intercomparison.
These results are generally consistent with recent
calculations of dust forcings at the surface and the TOA
over the North Atlantic (Evan et al. 2008; Myhre et al.
2003; Yoshioka et al. 2007). Our shortwave forcing is
also generally consistent with that reported by Carlson
and Benjamin (1980) and Zhu et al. (2007), although our
longwave forcing is smaller than the forcing reported in
either study.
On the basis of satellite observation of Saharan dust
over the Atlantic Ocean, Li et al. (2004) and Christopher
TABLE 4. Same as Table 3 but for the CAM3 dust. The columnintegrated dust AOT at 0.64 mm for each subregion and the whole
MDR is used for the calculations.
SWR
LWR
Dust AOT
TOA
TOA efficiency
Srf.
Srf. efficiency
TOA
TOA normalized
Srf.
Srf. normalized
Region
I
Region
II
Region
III
0.25
27.16
228.6
211.1
44.6
1.17
4.67
3.34
13.3
0.24
26.92
228.8
210.7
244.5
1.12
4.65
3.43
14.3
0.14
23.74
226.7
25.87
241.9
0.54
3.85
1.96
14.0
MDR
0.19
25.40
228.4
28.39
244.2
0.84
4.42
2.67
5156
JOURNAL OF CLIMATE
VOLUME 22
FIG. 5. Changes in net downward radiative fluxes (W m22) by the (a) MODIS–MATCH
dust, (b) CAM3 dust, (c) MODIS–MATCH SAL (both dust and dry anomaly), and (d) CAM3
SAL over region 1 for the cases of column-integrated AOT . 0.8 (see Table 1) for August–
September in 2003–06. The solid lines are the changes in total fluxes, the dashed lines are the
changes in shortwave fluxes, and the dashed–dotted lines are changes in longwave fluxes.
and Jones (2007) report diurnal-mean shortwave forcing efficiency at the TOA of 232 to 238 W m22/AOT
and 247.9 6 3.81 W m22/AOT in June–August, respectively. Our shortwave forcing efficiency at the TOA
for August–September over the MDR is 230.0 W m22/
AOT for the MODIS–MATCH dust and 228.4 W m22/
AOT for the CAM3 dust, close to the value of
–24.8 W m22/AOT reported in Yoshioka et al. (2007)
for June–August over the North Atlantic Ocean.
Christopher and Jones (2007) also reports an estimated longwave forcing efficiency of 8.96 6 3.51 W m22/
AOT at the TOA, larger than our value of 4.30 (4.42)
W m22/AOT for the MATCH-MODIS (CAM3) dust
and Yoshioka et al. (2007)’s value of 5.3 W m22/AOT.
We notice that our values of dust forcing are slightly
different from those in Yoshioka et al. (2007), although
we apply similar dust optical properties. The difference
between our and their results may arise from the difference in the period and region being studied as well as
the different radiative transfer models being used.
c. Radiative heating flux anomaly by the SAL
To estimate the total heating flux anomaly as a result
of the SAL, we have to include both the dust and the dry
anomaly. Figure 3 shows how the humidity of the SAL
varies with the total AOT. Thus, it can be used with
Table 1 to include both the dust and dry anomaly in the
three subregions in our radiative transfer calculation.
For example, for the SAL with the total AOT greater
than 0.8 in region I, we include the corresponding dry
anomaly (i.e., the thick-dashed line in Figs. 3a,d) together with the column-integrated dust AOT (0.98 or
0.78, see Table 1) in our heating flux anomaly calculation. The resulting changes in the daily averaged net
downward radiative fluxes for this SAL perturbation are
shown in Figs. 5c,d, together with the calculations including the dust-only in Figs. 5a,b for the MODIS–
MATCH and CAM3 profiles, respectively.
For the dust-only case (Figs. 5a,b), the net downward
shortwave flux is reduced by scattering and absorption,
and the net downward longwave flux increases because
less upward flux from the surface can reach above the
dust layer (;700–900 hPa) and more downward flux
is emitted below the layer. The pattern of flux changes
is similar for both MATCH-MODIS and CAM3 dust
profiles, with the CAM3 calculation having smaller
magnitude of shortwave flux changes because of fewer
small-size particles.
In the SAL, both dust and dry anomalies change the
radiative fluxes. Since water vapor is a greenhouse gas,
the removal of water vapor in the lower troposphere
reduces the greenhouse trapping of longwave radiation
1 OCTOBER 2009
WONG ET AL.
5157
FIG. 6. (a) Climatology of heating rate anomalies (K day21) caused by the MODIS–MATCH
dust of column-integrated AOT of one for August–September in 2003–06 over region 1. The
total heating rate anomaly (solid line) is decomposed into shortwave (dashed line) and longwave (dashed–dotted line) heating anomalies. Total heating rate anomalies (K day21) as
functions of column-integrated AOT for (b) region 1, (c) region 2, and (d) region 3.
by water vapor, in turn modifying the radiative flux
anomalies caused by dust. This is illustrated in Figs. 5c,d
(the dashed–dotted line). Compared to the pure dust
case, there is a decrease in the net downward longwave
flux above the SAL because less water vapor increases
the upward longwave radiation from the earth’s surface.
Below the SAL, there is also a decrease in the net
downward longwave flux (compared to the dust-only
case) because of the reduced emission from the atmospheric water vapor. Since CAM3 simulates larger dry
anomalies than the NCEP reanalysis (Fig. 3), the change
of longwave flux in the CAM3 SAL is larger.
d. Radiative heating rate anomaly by dust
and the SAL
The daily averaged heating anomaly can be calculated
from vertical convergence of the daily averaged heating
flux anomaly. Figures 6–9 show the dust and the SAL
radiative heating anomalies over the three regions in
the MDR.
Absorption of shortwave radiation (SWR) by the dust
heats the atmosphere around the dust layer (the dashed
line in Figs. 6a, 7a). CAM3 has a lower-altitude dust
layer and therefore the peaks of its shortwave heating
anomalies are located at a lower altitude. The dust layer
traps the longwave radiation (LWR) from the surface,
cooling the atmosphere at and above the dust peak and
warming the atmosphere below the peak (the dashed–
dotted line in Figs. 6a, 7a). The net heating rate, therefore, has a peak below the dust concentration peak
(;600–800 hPa for the MODIS–MATCH and ;800–
900 hPa for the CAM3). The heating anomaly maximum
is ;0.2–0.3 K day21, located between 850–900 hPa for
the MODIS–MATCH, and ;0.2–0.4 K day21, near the
surface for the CAM3. Carlson and Benjamin (1980)
estimated the peak of the net heating rate of about
0.8 K day21 located at 700 hPa. The difference between
our and their results stems from their lower singlescattering albedo and more large particles than in our
calculation (Yoshioka et al. 2007). Our heating rate
anomaly profiles are similar to that computed by Zhu
et al. (2007) for the Saharan coast, with most heating
located at altitudes below 600 hPa.
Figures 6b–d, 7b–d show the net dust heating anomalies for regions 1–3 as a function of AOT for the MODIS–
MATCH and the CAM3, respectively. The profiles of
these heating anomalies do not resemble those of the
temperature anomalies shown in Fig. 2. For example, the
peaks of the heating rates in Figs. 6, 7 are not located at
850 hPa, where the peaks in temperature anomalies in
Fig. 2 are located. Therefore, the radiative effect of dust
alone cannot explain the temperature anomaly.
5158
JOURNAL OF CLIMATE
VOLUME 22
FIG. 7. Same as Fig. 6 but for the heating rate anomalies caused by the CAM3 dust.
If we include the dry anomaly, the SAL heating
anomalies for AOT . 0.8 are ;0.1–0.4 K day21 at
850 hPa (Figs. 8b–d, 9b). The dry anomaly reduces the
longwave heating below 900 hPa compared to the pure
dust case (comparing the dashed–dotted line in Fig. 8a to
that in either Fig. 6a or Fig. 9a to that in Fig. 7a), because
of the less downward emission from water vapor, and
results in a maximum of heating at 850 hPa. The heating
by the SAL at altitudes below 600 hPa implies that the
dust and dry anomaly of the SAL together may help
maintain its temperature inversion during its transport
across the MDR.
FIG. 8. Same as Fig. 6 but for the heating rate anomalies caused by the MODIS–MATCH SAL
(both dust and dry anomaly).
1 OCTOBER 2009
WONG ET AL.
5159
FIG. 9. Same as Fig. 7 but for the heating rate anomalies caused by the CAM3 SAL (both dust
and dry anomaly).
There are drastic differences of the SAL heating rate
anomalies in the middle troposphere (;500 hPa) between the MODIS–MATCH and the CAM3. The CAM3
has positive heating rate anomalies around 500 hPa in
regions 1 and 2, whereas the MODIS–MATCH has
negligible anomalies there. Since radiative heating rate
anomalies caused by pure dust do not show these drastic
differences around 500 hPa (cf. Figs. 6, 7), differences in
the water vapor profiles between the CAM3 and the
NCEP dataset are likely the main cause of this difference.
e. Maintenance of the temperature inversion in the
lower troposphere
The atmosphere tends to relax any temperature
anomaly back to the background conditions through
thermal radiation. With the heating caused by the dust
and the dry anomalies in the lower troposphere, the
cooling generated by thermal relaxation is compensated. Table 5 shows the instantaneous total cooling rate
anomalies of the SAL (including dust, warm, and dry
anomalies) at the top of the inversion layer as well as the
thermal cooling rate anomalies (without dust and dry
anomalies) at the same altitude. The top of the inversion
layer is defined at the altitude of the maximum temperature anomaly in each case shown in Fig. 2: for the
MODIS–MATCH cases, it is at 850 hPa for regions 1
and 2 and at 700 hPa for region III; for the CAM3 cases,
it is at 850 hPa for almost all the cases but at 925 hPa for
the case of AOT . 0.8 in region I.
In all regions, the SAL helps maintain its inversion
layer by reducing thermal cooling rates. The thermal
cooling rates are reduced by more than a factor of 2 in
regions I and II for the MODIS–MATCH cases. In region
III, the MODIS–MATCH SAL begins to lose its ability
to maintain the thermal relaxation when AOT falls below
0.6. As shown in Fig. 8, the heating anomalies in 600–
850 hPa are almost zero when AOT falls below 0.6. Since
the heating anomalies from dust do not vary much with
regions, the change in the SAL’s ability to maintain heat
anomalies across the regions mainly arises from the
change in the water vapor profiles. In CAM3, the peaks of
water vapor anomalies are still located at 900 hPa for
AOT in 0.4–0.6; therefore, the model’s SAL in region III
still has the ability to maintain its inversion layer.
However, we also note that there is more uncertainty
in the temperature, water vapor, and dust profiles in
region III. The heating ability of the SAL in region III is
still subject to questions. Further research is necessary to
determine how dust and dry anomalies play a role in
thermal property of the SAL in the western Atlantic.
4. Conclusions and discussion
Although many previous efforts have focused on the
radiative effect of dust (Carlson and Benjamin 1980;
Evan et al. 2008; Li et al. 2004; Myhre et al. 2003; Weaver
et al. 2002; Yoshioka et al. 2007; Yu et al. 2006), this
work is the first to quantify the relative role of dust and
5160
JOURNAL OF CLIMATE
VOLUME 22
TABLE 5. The MODIS–MATCH SAL and CAM3 SAL total heating rates (K day21), including dust, dry, and warm anomalies at the
top of the inversion layer. The numbers in the parentheses are the heating rates without the dust and dry anomalies. The top of the
inversion layer is defined at the altitude of the maximum temperature anomaly in Fig. 2. For MODIS–MATCH, it is at 850 hPa for regions
I and II and at 700 hPa for region III. For CAM3 it is at 850 hPa for almost all cases, except for AOT . 0.8 in region I, it is at 925 hPa.
AOT
Region I
MODIS–MATCH
CAM3
Region II
MODIS–MATCH
CAM3
Region III
MODIS–MATCH
CAM3
.0.8
0.6–0.8
0.5–0.6
0.4–0.5
0.3–0.4
20.19 (20.52)
20.73 (21.09)
20.20 (20.46)
20.43 (20.72)
20.14 (20.36)
20.40 (20.63)
20.10 (20.28)
20.29 (20.47)
20.04 (20.18)
20.23 (20.33)
0.02 (20.25)
—
20.06 (20.24)
20.25 (20.60)
20.06 (20.20)
20.16 (20.42)
20.08 (20.18)
20.10 (20.25)
20.04 (20.11)
20.10 (20.14)
20.28 (20.34)
—
20.17 (20.22)
—
20.23 (20.21)
0.00 (20.39)
20.16 (20.14)
20.01 (20.18)
20.12 (20.10)
—
the associated dry anomalies in maintaining the atmospheric stability in the tropical cyclone main development region. Both dust and dry anomalies are important
in heating up the SAL in the lower troposphere, with
dust contributing about 50% of the maximum heating
rate anomalies when AOT . 0.8. East of 408W in the
MDR, the dust and dry anomalies reduce thermal
cooling at the top of the inversion layer and help
maintain the inversion layer. West of 408W, the SAL
may lose its ability to maintain the inversion layer for
AOT , 0.6. However, uncertainty in the dust, moisture,
and temperature profiles in this area necessitate further
investigation of this issue.
In the middle troposphere (;500 hPa), where dust
concentration is low, the SAL heating rate is sensitive
to the local dry anomaly. The dryer midtroposphere
therefore has a larger local heating rate anomaly. Thus,
the CAM3 SAL has larger heating rate anomalies at
500 hPa than the MODIS-MATCH SAL. This may
partially explain why the cold anomaly associated with
the SAL at 500 hPa in the NCEP reanalysis temperature
is much larger than that in the CAM3 simulation. The
origin of the cold anomaly needs to be addressed in future investigations. Other diabatic heating processes
may also play a role in generating this cold anomaly; for
example, the suppressed deep convection associated
with the SAL (Wong and Dessler 2005) may cause a
reduction in latent heat release.
In this study, we have assumed the particle shape of
dust is spherical, and we were able to use the Lorenz–
Mie theory to compute the dust optical properties. Most
recently, Yang et al. (2007) found that the nonsphericity
effect of dust is essential to determine shortwave radiance at the top of the atmosphere. Further investigations
are necessary to assess the sensitivity of the dust heating
rates to the morphologies of dust particles. Moreover,
our dust optical properties were derived from observations near the west coast of Africa. How dust optical
properties change during its transport across North Atlantic Ocean is still an open question that needs further
research. Finally, in this study, we did not consider
feedbacks (e.g., changes in vertical motion and cloud
cover) associated with the radiative heating anomaly. A
more thorough investigation that includes feedbacks
should involve experiments with a general circulation
model coupled with dust transport and radiative properties (e.g., Yoshioka et al. 2007).
This study has indicated that several aspects of the
SAL properties are missing in the CAM3 simulation.
First, the missing cold anomalies in the CAM3 SAL
imply that CAM3 simulates a dryer midtroposphere in
the SAL. Second, the missing lifting of the lower tropospheric warm and dry anomalies in the CAM3 SAL
while transported across the tropical Atlantic implies
that the elevation of the inversion layer may not be
appropriately simulated.
Given the discoveries of the negative trend of dust
loading over the tropical North Atlantic (Foltz and
McPhaden 2008; Wong et al. 2008), the effect of this
decreasing dust loading on the long-term variation of the
stability over the tropical North Atlantic requires further detailed investigations. As shown in this study,
water vapor profiles are as important as the dust loading
in assessing the heating rates in the dust layer. Therefore, the quality of current reanalysis, remote sensing
retrievals, and model simulations of water vapor over
dusty areas needs to be improved to accurately assess
the heating rates in the dust layer.
Acknowledgments. We thank the NASA MODIS
Science Team for the collection-5 product, NASA
CALIPSO–CloudSat Science Team for the aerosol layer
product, and their effort for making it easy to access the
data from the LAADS and ASDC Web sites. We also
thank Peter Colarco at NASA Goddard Space Flight
Center for his detailed discussion of the optical properties
1 OCTOBER 2009
WONG ET AL.
of Saharan dust and two anonymous reviewers for their
comments. This work is supported by NASA Grant
NNX07AR12G to Texas A&M University.
REFERENCES
Carlson, T. N., and J. M. Prospero, 1972: The large-scale movement
of Saharan air outbreaks over the northern equatorial Atlantic. J. Appl. Meteor., 11, 283–297.
——, and S. G. Benjamin, 1980: Radiative heating rates for Saharan dust. J. Atmos. Sci., 37, 193–213.
Chou, M.-D., and M. J. Suarez, 2002: A solar radiation parameterization for atmospheric studies. Tech. Rep. Series on
Global Modeling and Data Assimilation, NASA/TM-1999104606/VOL15, 38 pp.
——, ——, X.-Z. Liang, and M. M.-H. Yan, 2003: A thermal infrared radiation parameterization for atmospheric studies.
Tech. Rep. Series on Global Modeling and Data Assimilation,
NASA/TM-2001-104606/VOL19, 68 pp.
Christopher, S. A., and T. Jones, 2007: Satellite-based assessment
of cloud-free net radiative effect of dust aerosols over the
Atlantic Ocean. Geophys. Res. Lett., 34, L02810, doi:10.1029/
2006GL027783.
Colarco, P. R., O. B. Toon, O. Torres, and P. J. Rasch, 2002: Determining the UV imaginary index of refraction of Saharan
dust particles from Total Ozone Mapping Spectrometer data
using a three-dimensional model of dust transport. J. Geophys.
Res., 107, 4289, doi:10.1029/2001JD000903.
——, ——, and B. N. Holben, 2003: Saharan dust transport to the
Caribbean during PRIDE: 1. Influence of dust sources and
removal mechanisms on the timing and magnitude of downwind aerosol optical depth events from simulations of in situ
and remote sensing observations. J. Geophys. Res., 108, 8589,
doi:10.1029/2002JD002658.
Collins, W. D., and Coauthors, 2006: The formulation and atmospheric simulation of the Community Atmosphere Model
Version 3 (CAM3). J. Climate, 19, 2144–2161.
Dubovik, O., and Coauthors, 2002: Variability of absorption and
optical properties of key aerosol types observed in worldwide
locations. J. Atmos. Sci., 59, 1959–1966.
Dunion, J. P., and C. S. Velden, 2004: The impact of the Saharan air
layer on Atlantic tropical cyclone activity. Bull. Amer. Meteor.
Soc., 85, 353–365.
——, and C. S. Marron, 2008: A reexamination of the Jordan mean
tropical sounding based on awareness of the Saharan air layer:
Results from 2002. J. Climate, 21, 5242–5253.
Evan, A. T., J. Dunion, J. A. Foley, A. K. Heidinger, and
C. S. Velden, 2006: New evidence for a relationship between
Atlantic tropical cylone activity and African dust outbreaks.
Geophys. Res. Lett., 33, L19813, doi:10.1029/2006GL026408.
——, and Coauthors, 2008: Ocean temperature forcing by aerosols
across the Atlantic tropical cyclone development region.
Geochem. Geophys. Geosyst., 9, Q05V04, doi:10.1029/
2007GC001790.
Foltz, G. R., and M. J. McPhaden, 2008: Trends in Saharan dust and
tropical Atlantic climate during 1980–2006. Geophys. Res.
Lett., 35, L20706, doi:10.1029/2008GL035042.
Goldenberg, S. B., C. W. Landsea, A. M. Mestas-Nuñez, and
W. M. Gray, 2001: The recent increase in Atlantic hurricane
activity: Causes and implications. Science, 293, 474–479.
Hand, J., N. Mahowald, Y. Chen, R. Siefert, C. Luo,
A. Subramaniam, and I. Fung, 2004: Estimates of atmospheric-
5161
processed soluble iron from observations and a global mineral
aerosol model: Biogeochemical implications. J. Geophys. Res.,
109, D17205, doi:10.1029/2004JD004574.
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.
Kanamitusu, M., W. Ebisuzaki, J. Woollen, S. K. Yang, J. J. Hnilo,
M. Fiorino, and G. L. Potter, 2002: NCEP–DOE AMIP II
reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643.
Karyampudi, V. M., and T. N. Carlson, 1988: Analysis and numerical simulations of the Saharan air layer and its effect on
easterly wave disturbances. J. Atmos. Sci., 45, 3103–3136.
——, and Coauthors, 1999: Validation of the Saharan dust plume
conceptual model using lidar, Meteosat, and ECMWF data.
Bull. Amer. Meteor. Soc., 80, 1045–1075.
Kaufman, Y. J., D. Tanré, L. A. Remer, E. Vermote, A. Chu, and
B. N. Holben, 1997: Operational remote sensing of tropospheric
aerosol over land from EOS moderate resolution imaging
spectroradiometer. J. Geophys. Res., 102, 17 051–17 067.
——, ——, O. Dubovik, A. Karnieli, and L. A. Remer, 2001:
Absorption of sunlight by dust as inferred from satellite and
ground-based remote sensing. Geophys. Res. Lett., 28,
1479–1482.
——, I. Koren, L. A. Remer, D. Tanré, P. Ginoux, and S. Fan, 2005:
Dust transport and deposition observed from the TerraModerate Resolution Imaging Spectroradiometer (MODIS)
spacecraft over the Atlantic Ocean. J. Geophys. Res., 110,
D10S12, doi:10.1029/2003JD004436.
King, M. D., and Coauthors, 2003: Cloud and aerosol properties,
precipitable water, and profiles of temperature and water
vapor from MODIS. IEEE Trans. Geosci. Remote Sens., 41,
442–458.
Léon, J.-F., D. Tanré, J. Pelon, Y. J. Kaufman, J. M. Haywood, and
B. Chatenet, 2003: Profiling of a Saharan dust outbreak based
on a synergy between active and passive remote sensing.
J. Geophys. Res., 108, 8575, doi:10.1029/2002JD002774.
Levy, R. C., and Coauthors, 2003: Evaluation of the ModerateResolution Imaging Spectroradiometer (MODIS) retrievals
of dust aerosol over the ocean during PRIDE. J. Geophys.
Res., 108, 8594, doi:10.1029/2002JD002460.
Li, F., A. M. Vogelmann, and V. Ramanathan, 2004: Saharan dust
aerosol radiative forcing measured from space. J. Climate, 17,
2558–2571.
Luo, C., N. Mahowald, and J. del Corral, 2003: Sensitivity study of
meteorological parameters on mineral aerosol mobilization,
transport, and distribution. J. Geophys. Res., 108, 4447,
doi:10.1029/2003JD003483.
Mahowald, N., C. S. Zender, C. Luo, D. Savoie, O. Torres, and
J. del Corral, 2002: Understanding the 30-year Barbados
desert dust record. J. Geophys. Res., 107, 4561, doi:10.1029/
2002JD002097.
——, C. Luo, J. del Corral, and C. S. Zender, 2003: Interannual
variability in atmospheric mineral aerosols from a 22-year
model simulation and observational data. J. Geophys. Res.,
108, 4352, doi:10.1029/2002JD002821.
——, D. R. Muhs, S. Levis, P. J. Rasch, M. Yoshioka, C. S. Zender,
and C. Luo, 2006: Change in atmospheric mineral aerosols in
response to climate: Last glacial period, preindustrial, modern,
and doubled carbon dioxide climates. J. Geophys. Res., 111,
D10202, doi:10.1029/2005JD006653.
Myhre, G., A. Grini, J. M. Haywood, F. Stordal, B. Chatenet,
D. Tanré, J. K. Sundet, and I. S. A. Isaksen, 2003: Modelling
the radiative impact of mineral dust during the Saharan Dust
5162
JOURNAL OF CLIMATE
Experiment (SHADE) campaign. J. Geophys. Res., 108, 8679,
doi:10.1029/2002JD002566.
Nalli, N. R., and Coauthors, 2005: Profile observations of the Saharan air layer during AEROSE 2004. Geophys. Res. Lett., 32,
L05815, doi:10.1029/2004GL022028.
——, and Coauthors, 2006: Ship-based measurements for infrared sensor validation during Aerosol and Ocean Science Expedition 2004. J. Geophys. Res., 111, D09S04, doi:10.1029/
2005JD006385.
Patterson, E. M., 1981: Optical properties of the crustal aerosol:
Relation to chemical and physical characteristics. J. Geophys.
Res., 86, 3236–3246.
Prospero, J. M., and T. N. Carlson, 1972: Vertical and area distributions of Saharan dust over the western equatorial North
Atlantic Ocean. J. Geophys. Res., 77, 5255–5265.
Rasch, P. J., N. M. Mahowald, and B. E. Eaton, 1997: Representations of transport, convection, and hydrologic cycle in chemical
transport models: Implications for the modeling short-lived and
soluble species. J. Geophys. Res., 102, 28 127–28 138.
Reid, J. S., and Coauthors, 2003: Comparison of size and morphological measurements of coarse mode dust particles
from Africa. J. Geophys. Res., 108, 8593, doi:10.1029/
2002JD002485.
Remer, L. A., and Coauthors, 2002: Validation of MODIS aerosol
retrieval over ocean. Geophys. Res. Lett., 29, 8008, doi:10.1029/
2001GL013204.
Sinyuk, A., O. Torres, and O. Dubovik, 2003: Combined use of
satellite and surface observations to infer the imaginary part of
refractive index of Saharan dust. Geophys. Res. Lett., 30, 1081,
doi:10.1029/2002GL016189.
Sokolik, I., A. Andronova, and T. C. Johnson, 1993: Complex refractive index of atmospheric dust aerosols. Atmos. Environ.,
27A, 2495–2502.
Szczodrak, M. P., P. J. Minnett, N. R. Nalli, and W. F. Feltz, 2007:
Profiling the lower troposphere over the ocean with infrared
hyperspectral measurements of the Marine-Atmosphere
Emitted Radiance Interferometer. J. Atmos. Oceanic Technol.,
24, 390–401.
Volz, F. E., 1973: Infrared optical constants of ammonium sulfate,
Sahara dust, volcanic pumice, and flyash. Appl. Opt., 12, 564–568.
Weaver, C. J., P. Ginoux, N. C. Hsu, M.-D. Chou, and J. Joiner,
2002: Radiative forcing of Saharan dust: GOCART model
simulations compared with ERBE data. J. Atmos. Sci., 59,
736–747.
VOLUME 22
Winker, D. M., and Coauthors, 2003: The CALIPSO mission:
Spaceborne lidar for observation of aerosols and clouds. Lidar
Remote Sensing for Industry and Environment Monitoring III,
U. N. Singh, T. Itabe, and Z. Lui, Eds., International Society
for Optical Engineering (SPIE Proceedings, Vol. 4893), 4891–
4911.
Wong, S., and A. E. Dessler, 2005: Suppression of deep convection
over the tropical North Atlantic by the Saharan air layer.
Geophys. Res. Lett., 32, L09808, doi:10.1029/2004GL022295.
——, P. R. Colarco, and A. E. Dessler, 2006: Principal component
analysis of the evolution of the Saharan air layer and dust
transport: Comparison between a model simulation and
MODIS and AIRS retrievals. J. Geophys. Res., 111, D20109,
doi:10.1029/2006JD007093.
——, A. E. Dessler, N. M. Mahowald, P. R. Colarco, and A. da
Silva, 2008: Long-term variability in Saharan dust transport
and its link to North Atlantic sea surface temperature. Geophys. Res. Lett., 35, L07812, doi:10.1029/2007GL032297.
Yang, P., and Coauthors, 2007: Modeling of the scattering and
radiative properties of nonspherical dust-like aerosols.
J. Aerosol Sci., 38, 995–1014.
Yoshioka, M., N. M. Mahowald, A. J. Conley, W. D. Collins,
D. W. Fillmore, C. S. Zender, and D. B. Coleman, 2007: Impact of desert dust radiative forcing on Sahel precipitation:
Relative importance of dust compared to sea surface temperature variations, vegetation changes, and greenhouse gas
warming. J. Climate, 20, 1445–1467.
Yu, H., R. E. Dickinson, M. Chin, Y. J. Kaufman, B. N. Holben,
I. V. Geogdzhayev, and M. I. Mishchenko, 2004: Annual cycle
of global distributions of aerosol optical depth from integration of MODIS retrievals and GOCART model simulations.
J. Geophys. Res., 108, 4128, doi:10.1029/2002JD002717.
——, and Coauthors, 2006: Direct radiative effect of aerosols as
determined from a combination of MODIS retrievals and
GOCART simulations. J. Geophys. Res., 109, D03206,
doi:10.1029/2003JD003914.
Zender, C. S., B. Huisheng, and D. Newman, 2003: Mineral Dust
Entrainment and Deposition (DEAD) model: Description
and 1990s dust climatology. J. Geophys. Res., 108, 4416,
doi:10.1029/2002JD002775.
Zhu, A., V. Ramanathan, F. Li, and D. Kim, 2007: Dust plumes
over the Pacific, Indian, and Atlantic oceans: Climatology and
radiative impact. J. Geophys. Res., 112, D16208, doi:10.1029/
2007JD008427.