We present final and quality-assured results of multiwavelength
polarization/Raman lidar observations of the Saharan air layer (SAL) over the
tropical Atlantic. Observations were performed aboard the German research
vessel R/V
Dust particles can travel over long distances of more than
10 000
The Saharan desert is the world's largest mineral dust source
To investigate dust and its climate-relevant aspects, comprehensive dust field
experiments (ground-based and airborne activities, in situ measurements
combined with active and passive remote sensing) are required with focus on
the complex relationship between the microphysical, chemical, morphological
shape, optical, radiative, and cloud-process-relevant properties of dust
particles. The latest attempt to characterize dust over scales of several
thousands of kilometers of travel distance (equivalent to 5–10 days of
travel time) has been performed by the series of well-defined field
activities: the Saharan Mineral Dust Experiments – SAMUM-1 (southern Morocco,
summer 2006)
Winter-mode as well as summer-mode dust transport regimes
The main goal of the SAMUM and SALTRACE activities was to conduct a detailed
vertically resolved characterization of Saharan dust close to the source as
well as within the Saharan air layer (SAL) on the way towards the Caribbean,
more than 5000 km downwind of the main source regions. Well-designed efforts
of combined airborne and ground-based in situ aerosol observations and remote
sensing were realized during all of the three campaigns. To better link the
SAMUM and SALTRACE results, continuous lidar observations were conducted
aboard the German research vessel R/V
The fully automated multiwavelength polarization/Raman lidar
In a series of two articles, we now present the final results of the cruise.
In Part 1, we discuss the dust layering characteristics in large detail
(Sect. 3.2) and the aerosol optical properties of four cases (Sect. 3.3),
which we denote as key stages of the SAL evolution across the Atlantic. We
discuss the intermediate stages of dust transport in addition to the two
stages of dust layering shortly after emission and about 10 days after
long-range transport across the Atlantic. These two cases with fresh and aged
dust were already presented by
Large quantities of Saharan dust are transported across the tropical Atlantic
throughout the year but more abundantly in the summer months. These dust
outbreaks are mostly confined to a deep mixed layer, commonly referred to as
the Saharan air layer (SAL) that often extends to 5–6
As we will show in Sect. 3, the well-mixed SAL resides above a relatively
dust-free layer, which we denote as marine aerosol layer (MAL). As mentioned,
the MAL comprises the humid, convective MBL, the top of which is frequently
capped by cumulus clouds and the occasionally occurring residual, less
convective layer from the MBL top to the SAL base. This residual layer below
the SAL (or, if absent, the upper part of the MBL) is denoted as a trade wind
inversion layer in the conceptual model. Marine aerosol particles dominate
the backscattering and extinction properties in the MAL according to our
shipborne lidar measurements. Note that another way to divide the vertical
column up to the SAL base (or the trade wind inversion) is related to cloud
formation and occurrence. The layer up to the base of trade wind cumuli is
called the sub-cloud layer. The remaining layer from the cloud base to the
SAL base or trade wind inversion is denoted as the cloud layer
The dust transport takes usually 5–7 days across the Atlantic. While the SAL
base rises with distance from Africa (from about 1500 m (850 hPa) height in
the Cabo Verde region to about 2500
The strong temperature inversion at the base of the SAL limits convective
activity within the SAL and consequently precludes the possibility of strong
wet removal from the dust layer (scavenging of dust particles below and
within clouds and removal by wash and rain out), except during periods with
deep convection and precipitation.
According to the modeling study of
The following questions now arise: are these features of the dust transport
across the Atlantic as described by the conceptual model in agreement with
our shipborne lidar observations? Are there aspects that are not described
and/or considered properly but have an impact on dust transport and removal?
The discussion is presented in Sect. 3.2 and 3.3. Before we present the
results in Sect. 3, we briefly describe the R/V
The first vertically resolved lidar study of the SAL across the tropical
Atlantic was presented by
However, all of these lidar studies are based on observations with standard
elastic-backscatter lidars which do not have the potential to directly
measure the particle extinction coefficient in the SAL and thus of the SAL
aerosol optical thickness (AOT). One of the most important input profiles in
the retrieval of the basic lidar products with these standard lidars is the
height profile of the particle extinction-to-backscatter coefficient (lidar
ratio). This input profile has to be estimated and can be very erroneous. In
the case of LITE, the dust lidar ratio was assumed to be in the range of 20
to 35
The transatlantic cruise M96 of the German R/V
The multiwavelength Raman/polarization lidar system Polly
One of the most important lidar-derived aerosol parameters in dust
observations is the volume depolarization ratio. This quantity is almost
directly measured. The volume linear depolarization ratio is obtained from
the calibrated ratio of the cross- to co-polarized backscatter signal
The particle depolarization ratio allows us to separate fine dust (dust
particles with diameters
Part of the results on the aerosol mixing state in the SAL are already
discussed here (in Part 1) so that the following sources of uncertainty need
to be mentioned. In discussions, frequently the question arises to what
extent dust aging by chemical processes, activation of dust particles to
serve as cloud condensation nuclei in cloud formation events, and by
size-dependent gravitational settling during long-range transport affect the
accuracy of the separation of dust from non-dust aerosol profiles. The
assumption of a universal dust depolarization ratio of about 0.3 may not be
generally valid. Aging of dust particles caused by cloud evolution and
subsequent evaporation processes or by chemical reactions on the dust
particle surface
Internal mixing of dust and non-dust particles may cause a misinterpretation
of the dust and non-dust particle extinction coefficients and mass
concentrations retrieved with the POLIPHON method. The internally mixed
non-dust aerosol contributions will be counted as dust contributions.
Chemical aging and successive water uptake at high humidity may significantly
change the overall (dust and non-dust) particle extinction coefficient
Regarding potential changes of the particle optical properties,
The lidar profile observations were accompanied by sun photometer
measurements in the framework of the Maritime Aerosol Network (MAN) as part
of the Aerosol Robotic Network (AERONET)
Radiosondes for measuring temperature, humidity, and wind profiles were
regularly launched at 12:00 and 00:00 UTC by the German Weather Service
aboard the ship
Figure
Figure
Three phases of dust transport are visible in Fig.
Marine and dust layers over the tropical Atlantic about
1000 km (case 1), 1700 km (case 2), 3300 km (case 3), and
4300 km (case 4) west of the African coast. The 532 nm range-corrected
signal
We selected four cases to study the evolution of the SAL and changes in the
dust optical properties with increasing distance from Africa; they are
indicated by numbers 1–4 in Fig.
The most remarkable feature in Fig.
To better understand the observed aerosol layering structures and vertical
exchange processes over the tropical ocean and to check the consistency with
the conceptual model in Sect. 3.2.1, we introduced the three layers (MBL, MAL, and SAL). The top
of the MBL is determined from the range-corrected signal profiles (shown in
Fig.
As was often observed during fair weather conditions (during SALTRACE over
Barbados, during SAMUM-2 over Cabo Verde), cumulus convection can intensify
and then the vertical extent of the clouds can increase from
300–500
To facilitate the comparison between the lidar observations in
Fig.
Wet deposition during times with clouds and precipitation as well as dry deposition contribute to the dust removal from the MAL. We analyzed METEOSAT satellite observations for the presence of strong cumulus convection and found that, except for case 4, wet deposition by deep convection and associated rain can be excluded. However, fair weather cumulus convection and light precipitation always occur over the tropical Atlantic, and thus a certain contribution of wet deposition to dust removal must be always taken into account.
Our observations are to a large extent in good agreement with the conceptual model. As already mentioned in the foregoing subsection, the observed changes in the MAL top height (increase), SAL base height and top height (decrease), and SAL vertical extent (decrease) with distance from Africa are similar to the ones described in the conceptual model. The observed sharp increase of the volume depolarization ratio at the interface between MAL and SAL suggests that injection of particles from the MAL into the SAL over the open Atlantic is almost impossible. Furthermore, the rather low depolarization ratio from the ocean surface to the MAL top suggests a fast and efficient removal of dust from the MAL by the vertical exchange processes.
It is noteworthy to mention that less sharp, more smooth structures in the
depolarization ratio at the MAL/SAL interface were observed over the western
part of Barbados during the SALTRACE summer campaigns in June and July 2013
and 2014
Gravitational settling plays the dominant role in the downward transport of
dust in the SAL over the open Atlantic according to the conceptual model.
However, our detailed profile observations presented in Sect. 3.3 as well as
the SALTRACE observations in Barbados in June and July 2013 and 2014
Profiles of the particle backscatter coefficient at 355, 532, and
1064 nm; extinction coefficient, extinction-to-backscatter ratio (lidar
ratio), and particle linear depolarization ratio at 355 and 532 nm; and
backscatter-related (bsc) and extinction-related (ext) Ångström
exponents. Temperature and relative humidity from radiosondes (cases 2–4) and GDAS
data (case 1) are given in addition. Mean profiles of the
optical properties for the time periods on 23 May 2013, 03:45–05:00 UTC
(case 1); 14–15 May 2013, 23:45–00:15 UTC (case 2); 9 May 2013,
23:15–24:00 UTC (case 3); and 5 May 2013, 23:40–24:00 UTC (case 4) are
shown. The label 532 N/R denotes the 532 nm near-range receiver channel. The
vertical signal smoothing length for the profiles of backscatter coefficient
and particle linear depolarization ratio is 457.5 m; the rest is smoothed
with 562.5 m window length. The signal-averaging periods are indicated by
white vertical lines in Fig.
The 5-day to 13-day HYSPLIT backward trajectories for 23 May 2013, 04:00 UTC (case 1), 15 May 2013, 00:00 UTC (case 2), 9 May 2013, 23:00 UTC (case 3), and 5 May 2013, 23:00 UTC (case 4). Symbols indicate air mass transport from day to day. The arrival height level of 500 m (red) is in the MAL. Arrival heights of 1500–3000 m (blue, green) are in the SAL. In addition, fires (red dots) detected by MODIS aboard the Terra and Aqua satellites are shown accumulated over a 10-day period (21–30 April 2013 for cases 3 and 4, 11–20 May 2013 for case 2, and 21–30 May 2013 for case 1).
In Fig.
A strong dust outbreak was observed on 23 May in Cabo Verde (case 1).
According to the backward trajectories in Fig.
The backward trajectories for case 2 (15 May 2013, 00:00 UTC) suggest a
possible impact of smoke in the upper half of the SAL, i.e., above 2 km
height. The trajectory for the arrival height of 2.5 km crossed fire areas
at heights well within the continental boundary layer. Fire smoke uptake was
possible during almost 2 days. The contribution of African smoke and haze to
particle extinction at 532
Compared to case 1, the daily mean AERONET Ångström exponent shows slightly enhanced values of 0.3 which may be an indication of the presence of an external mixture of dust and fine-mode particles of continental origin but may also show the increasing influence of marine aerosols in the MAL on the AOT with increasing distance from Africa.
The dust layer on 9–10 May (case 3) also potentially contained smoke,
according to the backward trajectory for the SAL center height of
2.5
The aged dust plume observed on 5 May (case 4) monotonically descended from
heights above 4500
As can be seen in Fig.
Dust-related particle extinction coefficients in the SAL of
40–80
The 532 nm particle depolarization ratios of
A critical point in our lidar data analysis is the smoke contribution to
backscattering and extinction.
In the discussion of the non-dust contributions to the SAL backscatter and
extinction coefficients, the influence of marine particles causing 532 nm
lidar ratios of 20–25
The simultaneous observations of depolarization ratios at 355 and
532
The backscatter- and extinction-related Ångström exponents in the SAL in
Fig.
Although the backscatter-related Ångström exponent is usually
The radiosonde profiles of temperature and relative humidity (RH) in
Fig.
Figure
Layer mean values of
The findings in Fig.
Our shipborne observations of the SAL mean extinction coefficients are in
good agreement with the SAMUM-2 observations in Cabo Verde (4 weeks in
May–June 2008) and the SALTRACE observations
As in the case of the SAL extinction coefficients, the layer mean lidar
ratios (on average,
The ship cruise allowed us also to describe clean marine conditions in terms
of lidar-specific optical properties at sites far away from continents. The
MAL lidar ratios (10–25
As mentioned at the end of Sect. 3.2.1 and emphasized in
Fig.
The question is now what processes can weaken the gravitational settling
effect.
Our observations are in accordance with these findings. If particle
sedimentation would be dominating in the SAL, we should observe a decrease of
the coarse dust fraction with height, i.e., an accumulation of the larger
dust particles in the lower part of the SAL after dust transport over days
and distances of 4000
Furthermore, the measured particle extinction coefficients range from 50 to
100
During a 1-month transatlantic cruise from Guadeloupe to Cabo
Verde (over 4500
The SAL vertical mean depolarization of
The shipborne lidar observations in May 2013 fit well into the dust
characteristics and layering structures gained from the SAMUM and SALTRACE
field campaigns regarding the long-range transport of dust. Good agreement
regarding the dust optical properties was found when comparing the dust
measurements in Morocco, Cabo Verde, and Barbados, and aboard the
R/V
The ship cruise also provided ideal conditions for the measurement of pure
marine aerosol optical properties far away from disturbing continents. The
results are consistent with the SAMUM-2 observations in Cabo Verde and the
winter SALTRACE observations in Barbados. During both campaigns, a few days
with pure marine conditions occurred. During the ship cruise, marine,
dust-free conditions in the lowest 1000–1500
In the companion paper of
The observations in this Part 1 and in the follow-up article (Part 2) as well as all the advanced lidar observations performed during the dust-related field campaigns in the last 10–15 years clearly indicate the importance of and need for comprehensive, vertically resolved dust measurements to better understand the life cycle of atmospheric dust, and the need to improve atmospheric dust modeling from emission to deposition, the interaction of dust with the radiation field, and the dust impact on cloud formation and precipitation. The buildup of a permanent ground-based dust-monitoring, lidar-networking infrastructure would be a big step forward to support the dust life cycle and impact research and dust forecasting by combining the continuously available lidar observations with dust modeling efforts.
To support and confirm our lidar-based findings (obtained by using many
assumptions), east–west long-distance research flights within the SAL across
the tropical Atlantic would be very helpful with the focus of the airborne in
situ aerosol observations on the aerosol mixing state, fractions of
internally and externally mixed aerosols, dust aging caused by cloud
processing and chemical aging, changes in the dust size distribution between
the Saharan dust source and the Caribbean, South and North America, and also
aging of biomass burning smoke particles in the dust–smoke mixtures during
the biomass burning season. Research flights are required during both summer
(dust) and winter (dust and smoke) half years. Further points of airborne
research should cover aerosol–cloud interaction aspects, i.e., cloud
condensation nuclei (CCN) and ice nucleating particle (INP) characterizations
across the Atlantic with the specific question regarding the contribution of
dust and non-dust particles to the CCN and INP reservoirs and how dust aging
changes the ability of dust to serve as CCN and/or INPs. The first steps in
this complex dust research field were performed during the SALTRACE campaign
Radiosondes for measuring, temperature, humidity, and wind
profiles were regularly launched aboard the research vessel at 12:00 and
00:00 UTC by the German Weather Service
The authors declare that they have no conflict of interest.
This article is part of the special issue “The Saharan Aerosol Long-range Transport and Aerosol-Cloud-interaction Experiment (SALTRACE) (ACP/AMT inter-journal SI)”. It does not belong to a conference.
We thank the R/V