Sahara dust storms during March 2004 have attracted much attention from the dust-research community due to their intensity, wide coverage, and endurance. In the present work, the dry deposition mechanisms of mineral dust are analysed during an event on the 3 March 2004 over the Northwest African coast. This particular case was chosen based on the strong dry removal that occurred, rendering it ideal for examining the deposition processes. The simulation of synoptic conditions and dry deposition of four dust particles including clay, small silt, large silt, and sand was performed with Eta model, coupled with a desert dust cycle module. The results have been compared with surface data from weather stations in North Africa, data of dry metals from stations located in Gran Canaria, and various satellite images such as European Organization for the Exploitation of Meteorological Satellites and Moderate Resolution Imaging Spectroradiometer for the period in question.
The global source strength of mineral aerosol is currently estimated to a value between 1000 and 5000 Mt·yr−1, for example, [
The presence of some metals such as Al, Fe, Co, and Mn, in dust may significantly modify the marine biochemistry and may change the phytoplankton communities resulting in fast growth rates leading to blooms [
Dry deposition is the transfer of airborne dust particles to the surface through different mechanical processes, mainly Brownian diffusion, interception, inertial impaction, and sedimentation [
Dry deposition is caused by impaction when an aerosol particle transported by the flow towards an obstacle cannot, when its inertia is too large, follow the flow deviation in the vicinity of the obstacle. Thus the particle collides with the obstacle surface and remains on its surface when the particle rebound is ignored.
Dry deposition also occurs via interception when particles of small inertia, which perfectly follow the streamlines of the mean flow field, pass in the vicinity of an obstacle and are held back because the distance between the particle centre and the surface is smaller than its radius [
The Brownian diffusion affects very fine particles, typically smaller than 0.1
The dry deposition velocities are calculated as a function of particle size, density, friction velocity, and surface characteristics and they include the contributions of turbulent transfer, Brownian diffusion, impaction, interception, gravitational settling, and particle rebound [
For the smallest clay and silt particles as shown in Table
Feature of typical dust particle.
|
Type | Typical particle |
Particle density |
Total mass |
---|---|---|---|---|
1 | Clay | 0.73 | 2.50 | 0.08 |
2 | Small silt | 6.10 | 2.65 | 1.00 |
3 | Large silt | 18.00 | 2.65 | 1.00 |
4 | Sand | 38.00 | 2.65 | 0.12 |
According to models [
This work focuses on the process of dry deposition during the event of 3 March 2004 over the African West coast when large quantities of dust are carried out of North Western Africa, in particular from Saharan source regions. Sahara dust storms during March 2004 have attracted much attention from the dust-research community due to their intensity, wide coverage, and endurance [
Due to the lack of surface observations across the Sahara, a rigorous quantitative verification of the model simulations is not possible. However the simulated dust patterns are compared with data from surface weather stations in North Africa, from dry metals from stations located in Gran Canaria (Spain) and various satellite images such as (EUMETSAT) M-et 8 and MODIS.
In order to investigate physical processes of dry deposition of the mineral dust over North West African coast, during an event on the 3 March 2004, the Eta system coupled with the module describing the dust cycle is applied [
Area of simulation.
The dust cycle module described by a set of
The dust module include four particles (
The size range of four soil particles classes [
The mass of the clay particles are estimated to be 1 to 2 orders of magnitude smaller than particles in range 1–10
The radius of silt particles are in the range from 1 to 25
A particle dry deposition parameterization scheme [
The dry deposition velocity
The particles properties and depositing surfaces (roughness, texture, and vegetation coverage) are characterized by the function
The most outstanding feature during the initial stages of this episode was the intensification of a high pressure system over the subtropical Atlantic Ocean accompanied by a cyclogenesis over central north Sahara.
On 3 March 2004 at 06 UTC, wind at 10 meters high struck the western regions of North Africa such as west Sahara, Northern Mali and Mauritania, and southwest Algeria. After that, it took an anticlockwise direction and then moved in a clockwise direction above the Algerian Sahara, Mali, and Mauritania. In south Algeria and west Libya, it gained speed remarkably ranging 10-11 ms−1. However, it decreased to 6 ms−1 over the areas of south Mali. In Mauritania, the wind ranged 3-4 ms−1 at the center and 6-7 ms−1 at the south. It weakened further ranging 3–5 ms−1 over the center of West Sahara and 4–7 ms−1 over the West African coast and near the surface of Canary Island.
At 12 UTC, wind at 10 meters high dominated all North African regions; it blew strongly over southwest Algerian Sahara, North Mali, Mauritania, and West Sahara reaching a value between 12 and 13 ms−1 over these regions. This increase in wind velocity was due to warm air masses at the center of southeast Europe, which moved in a clockwise direction reaching Northwest Africa. It continued through two ways: one headed towards southwest of North Africa and the second deviated through the Atlantic Ocean along southwest Europe.
Six hours later, It was clearly seen that near the surface the weakness wind at 10 meters high covered approximately the West parts of North Africa including the regions of southwest Algeria and Northern Mali, Mauritania,and West Sahara; it reached 6 to 7 ms−1; therefore these meteorological controls (the diminishing of wind velocity) establish the dust content of the atmosphere, hence its transpire dust deposition over these regions. The dust clouds from West Africa Figure
The island’s proximity to the West Saharan Desert results in the common presence of Saharan winds, 30% of the year, for a 5-year period [
According to these synoptic conditions, mineral dust transported produces frequent dry hazes over West Africa and West Africa coast on 3 March 2004 at 18 UTC. The dynamics of the synoptic conditions were described by [
Wind at 10 meters high on 3 March 2004 at (a) 06 UTC, (b) 12 UTC, and (c) 18 UTC.
Total dry deposition at 06 UTC in Figure
Dry deposition on 3 March 2004 at 06 UTC, (a) total dry deposition, (b) clay particle, (c) small silt particle, (d) large silt particle, and (e) sand particle. The black circle indicates the maxima of dry deposition.
Dry deposition of clay particles seem over the regions of Libya, Algeria Sahara, Mali, and Mauritania, also over West Sahara; it ended over the Atlantic Ocean; the greater value was distinct at south Algeria, Figure
The coarse large silt and sand particles, respectively, have a high settling velocity from the air and therefore it is not carried more than few hundred kilometers away by winds, as shown in Figures
At 12 UTC, total dry deposition expand more than the earlier time above West Africa, in the Sahara regions of south Algeria, Mali, and Mauritania, also south Morocco and West Sahara. The highest quantity evident ended in Mauritania Sahara is observed in Figure
Dry deposition on 3 March 2004 at 12 UTC, (a) total dry deposition, (b) clay particle, (c) small silt particle, (d) large silt particle, and (e) sand particle.
Figure
At this time the small silt particles have level amount in deposition than the other particles as shown in Figures
Regarding the large silt and sand particles, there are no height accumulations of dry deposition over Atlantic water and Canary Island, for the reason that the coarse masses of these particles as illustrated in Figures
It can be concluded that the small silt particle has greatest amount in dry deposition than the other particles on 3 March 2004 at 12 UTC. The meteorological conditions favors to uptake particle from the atmosphere and depending on the physical and chemical properties of their sizes during the same period, the large silt and sand set down closed to the source of dust production.
Total deposition for the period of 3 March 2004 at 18 UTC indicated the increase of dustiness in the regions of south Algeria, Sahara of Mali and also the regions of North Mauritania and West Sahara; the deposition also exposed over the Atlantic Ocean and Canary Island, Figure
Dry deposition on 3 March 2004 at 18 UTC, (a) total dry deposition, (b) clay particle, (c) small silt particle, (d) large silt particle, and (e) sand particle.
In Figure
Figure
At 18 UTC, as shown in Figure
For large silt and sand particles, gravitational settling becomes much more important and deposition velocity increases with increasing particle diameter, large particles fall, reaching a terminal velocity.
Removal processes of atmospheric mineral particles by dry deposition occurs at the surface in the vicinity of the dust source areas, and because larger particles cannot participate in long range transport, the deposition of clay and small silt particles are commonly investigated in Island of Grand Canaria from 1 to 3 March 2004.
In the model, in a layer close to the surface of 10 meters depth, known as the surface layer, the dry deposition dominates; this layer consists in two virtual layers usually considered to calculate dry deposition velocities [
The dry deposition velocity
Evolution of dry deposition velocity of both particles clay and small silt over Gran Canaria from 1 to 3 March 2004.
Firstly, for small silt particles having diameters of 12.2
Secondly, for particle sizes at 1.46
Dry deposition velocity does not only depend on particle properties, but also on atmospheric conditions and surface characteristics [
Depending on the mechanisms for the generation of turbulence [
The dry deposition event is treated in the surface layer because the entrainment of soil particles is determined by the momentum transfer from this layer to the surface and the motion of sand particles confined to this layer [
Evolution of turbulent deposition velocity of both clay and small silt particles between the layer
Evolution of turbulent deposition velocity at the top of viscous sublayer
In the model calculation, the surfaces with turbulent regimes ranging from smooth to rough conditions include sea, bare soil and ice surface are considered; however the dry deposition over the surface covered by vegetation according to scheme of [
In the two-layer dry deposition model, in the upper layer, turbulent diffusion dominates over molecular diffusion, and in lower layer the molecular diffusion dominates over turbulent diffusion [
In Figure
Thus the dry deposition velocity is a function of both turbulent velocities
It appears for all velocities from Figures
The simulation results are validated as far as observational data would permit. The data available for the comparison are the following.
The distributions of the surface weather stations in North Africa and the dust records for March 2004 are shown in Figure
Distribution of the surface weather stations in North Africa (black dots) and daily dust weather records for the period 3 March 2004 (a). The dust weather records for the entire month of March 2004 are shown in (b), (shaded) and the dust weather records (symbols) for 3 March 2004 [
Dust storms and severe dust storms which are shown in Figure
The results of dust dry deposition pattern and evolution are compared with the dust weather records in Figure
The results revealed that on 3 March 2004 dust storms and their deposition developed in northwestern Sahara in conjunction with the cyclogenesis and the formation of the cold front. Much of the dust emitted from Western Sahara was trapped in the cold air mass, forming a marked dust frontal structure [
The
Data from meteorological stations, namely,
From data in Figure
Comparison of wind speed at 10 meters high of (OT) station with the model results from 1 to 3 March 2004. The observational (blue graph) and simulated wind speed data (red graph) showed a good correlation score of decreasing wind velocity on 3 March 2004.
According to Table
On the graph of [
Settling velocity as a function of particle sizes.
(a) Dust clouds from West Africa extending to Atlantic Ocean and Canary Islands, detected by (EUMETSAT) Met-8 on 3 March 2004 at 12 UTC. (b) Dust storms from the Northwest Africa and high dust concentrations above Canary Island at 18 UTC where dry deposition occurred. The reddish rectangle indicates the occurring zone of dry deposition. (c) The Zoom image above the Canary Island showing high dust concentration originating from North West Africa regions. (a), (b), and (c) are taken from Visible Earth (2004).
According to the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) Met-8 and Moderate Resolution Imaging Spectroradiometer (MODIS) images for 3 March 2004, a cold air outbreak from Europe to Western Africa caused a major dust storm over large parts of West Africa. The dust was blown out across the Atlantic Ocean and out over Canary Islands as shown in Figure
Moderate Resolution Imaging Spectroradiometer detected Saharan dust storm of Northwest Africa above Canary Islands. (a) MODIS image with 2000 m resolution and (b) MODIS image with 250 m resolution. Images were taken from Visible Earth (2004). (c) Desert dust of West Africa out over Canary Island during 3 March 2004, photo: “Eugenio Rodriguez 03 March (2004)” and (d) dust over coast Gran Canaria, may be associated with Saharan dust storm of North Western Africa [
At 18 UTC, a thickest dry deposition over Mali and Mauritania, a border of West Sahara, as depicts in Figures
A comparison of Figures
The dry deposition during the event of 3 March 2004 can be qualitatively compared with the MODIS (on board NASA’s Aqua) images, as shown in Figure
African dust transport constitutes a large fraction of the annual atmospheric deposition in the Canary Islands. The analyses of aerosol samples and deposition measurements over Gran Canaria during different dust deposition episodes have been carried out during 3 March 2004 [
Dust sampling was carried out at three sample stations located in Gran Canaria: Pico de la Gorra (1930 m a.s.l., 27°56′N, 15°33′W), Tafira (269 m a.s.l, 28°06′N, 15°24′W), and Taliarte (close to sea level; 27°59.5′N, 15°22′W) as seen in Figure
Location of the three sampling sites in Gran Canaria: Pico de la Gorra (1930 m a.s.l., 27°56′N, 15°33′W), Tafira (269 m a.s.l, 28°06′N, 15°24′W), and Taliarte (close to sea level; 27°59.5′N, 15°22′W) [
Dry metal deposition fluxes of (a) Al, (b) Fe, (c) Mn, and (d) Co at Gran Canaria, high dry deposition marked on 3 March 2004 [
The uplifted dust has a lower grain size that the parent material is enriched in the clay fraction, which produces a mineralogical and then chemical fractionation [
Guieu et al. [
Mean concentration of Al included in mineral Sahara dust according to [
Mean concentration | |||
---|---|---|---|
Saharan dust | Saharan soils | Saharan dust end member | |
Al, % | 7.09 ± 0.79 (11%) | 5.78 ± 1.68 (29%) | 7.09 ± 0.79 (11%) |
In Figure
Concentration of Al in different dust episodes from Gran Canaria Stations converted to mineral dust concentration and compared with model simulation of dust deposition event of 3 March 2004.
The greatest differences between the mean and max concentrations of Al were observed in 3 March 2004. The Al concentration was increased by mineral dust transport with cyclones originating from North Africa [
The aluminum concentration in Gran Canaria appears to be a useful tracer and used as indicator for the intrusion of dust from surrounding desert of Northwest Africa areas, which are primary sources of mineral aerosol. According to Figure
As a result of the large Northwest gradient in the dust concentration in Gran Canaria, small shifts in the large scale wind systems or in the dust sources in Africa could result in very large changes in dust transport and in the related deposition to the areas of Canary Island. This applies to the day-to-day changes in dust concentrations (which are subject to the winds associated with the controlling large-scale meteorological situation) and also to longer term concentrations (which are related to climatologically factors and the associated long-term changes in meteorology).
According to [
Comparison between the polymodal textures fraction form the Aeolian dust samples in the Island of Gran Canaria and the model results of dry deposition fraction on 3 March 2004 at 18 UTC.
There were an increase in the very fine-fine and coarse silt fraction and a decrease of the coarse silt fraction, in agreement with conditions of suspension transport prior to accumulation. However, the particles chosen in this episode of dry deposition in Island of Gran Canaria are corresponding to those used in [
Comparison between textural analyses of Aeolian dust sample collected in Gran Canaria and model output related to fraction of dry deposition of four grain sizes.
Kubilay et al. [ |
Present work |
---|---|
Clay (<2 |
Clay (1,46 |
m. silt (8–16 |
Small silt (12,2 |
c.-v.c. silt (16–62 |
Large silt (36 |
v.f. sand (62–125 |
Sand (76 |
The model output related to fraction of dry deposition particles originating from airborne Saharan dust of Northwest Africa (Figure
The clay particles from the model show high deposition similar to the same particles collected in Gran Canaria, the decreasing of the percentage amount of this particle signifies that this particle can travel a long distances due to its small size and due to meteorological conditions favorable before it entrained in the deposition process in the Island of Gran Canaria. Due to high dry deposition of clay particles at 18 UTC, the small fraction of this particle shown by the model (30%) is equal to the maximum clay particles collected In Gran Canaria. It was concluded from these comparisons between the model results and the textural analyses of aeolian dust samples on 3 March 2004 at 18 UTC that a very important quantity of dry deposition from Northwest Africa Sahara was recorded. The dry deposition of the four particles over Gran Canaria follow to many factors such as particles diameter and the meteorological conditions associated with the aeolian dust.
The major sources for long-range transport of mineral dust in North-eastern subtropical Atlantic are located in arid regions of Northwest Africa. A long front of Saharan dust sweeps across south Algeria, Mali, Mauritania, and Western Sahara and produces frequent dry haze over the vicinity of the source areas; hence the deposition of clay and small silt particles is commonly investigated.
Throughout the event studied here, the dry deposition amount is governed by the physical process of dry deposition concurrent with wind at 10 meters above the ground; consequently most of the wind speeds tend to support evidence of an excellent link between the model wind velocity and (OT) station data.
The uplifted dust has lower grain sizes of the parent material including small silt and clay is a better reference of aerosol compositions for the eroded areas; hence the Al concentration taking place in the area of Gran Canaria was increased on 3 March 2004 when large quantities of dust were carried out of North Western Africa. A good correlation was found between the model simulation of mineral dust concentration and concentrations of Al data measured. The aluminum concentration in Gran Canaria appears to be a useful tracer and is used as an indicator for the intrusion of dust from the surrounding desert areas of Northwest Africa.
Grain size distribution curves were found to be Polymodal in the comparison between textural analyses of aeolian dust samples collected in Gran Canaria and model outputs related to the fraction of dry deposition of four particles including clay, small silt, large silt, and sand. At 18 UTC there was a significant quantity of dry deposition amount of both small silt and clay particles. However large particles fall, reaching a terminal velocity, and are not carried more than few hundred kilometers away.
Even though dry deposition velocity is depending not only on particle properties, but also on atmospheric conditions and surface characteristics, for small silt particles the deposition velocities rise steadily due to gravity and especially impaction and interception processes at the surface and by decreasing wind speeds; for particles of this size the deposition velocity increases with particle diameter as gravitational settling becomes more important, whereas clay particles are controlled by the turbulent processes (size independent); at the top of viscous layer, very close to the smooth surface, turbulence is generally very weak due to the very strong effect of viscous dissipation. For these sizes dry deposition does not appear to be an efficient removal processes, and this could lead to longer atmospheric residence times. For large silt and sand particles gravitational settling becomes much more important and deposition velocities increase with increasing particle diameter.
Dry deposition provides a significant mechanism for the removal of the four particles from the atmosphere, following many factors such as particle diameter and meteorological conditions associated with the aeolian dust and by a reason of the synoptic condition frequent occurrence of falling wind at 10 meters high, the dust deposition of both particles clay and small silt over the West Africa coast become extensive at 18 UTC on 3 March 2004.
As a result, the atmosphere over Western Africa is almost permanently dry with significant amount of mineral dust deposited during the event of 3 March 2004, from that the dust production is enhanced during this period because of the breaking of soil crusts, in this respect if the ground is drier more dust will rise and arrive to Gran Canaria. It is reasonable to assume that the factors affecting the Gran Canaria dust dry deposition record depend on particle size structure and by the occurrence of proper synoptic conditions.
The authors declare that there is no conflict of interests regarding the publication of this paper.
The authors thank the group of Atmospheric Modeling and Weather Forecasting of the University of Athens. A special acknowledgment is to the University of de Las Palmas de Gran Canaria (Spain), the Instituto de Astrofisica de Canarias (IAC), the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT), and Moderate Resolution Imaging Spectroradiometer (MODIS) for providing them measured data and satellite Images.