Adam A. Atia 1,2, Vernon Morris3, and Reginald Blake4

OBSERVING AEROSOL MASS DENSITIES DURING THE TRANS-ATLANTIC TRANSPORT OF SAHARAN DUST AND BIOMASS BURNING AEROSOLS

1NOAA-CREST Research Experiences for Undergraduates Program

and

2The City College of New York

3Chemistry and Atmospheric Sciences, NOAA Center for Atmospheric Sciences

Howard University, Washington, D.C.

4Department of Physics

New York City College of Technology, Brooklyn, New York

ABSTRACT

Saharan dust is transported by the Northeast Trade Winds from Northern Africa to the Caribbean, South America, and eastern seaboard of the US, generally affecting atmospheric chemistry, oceanic bioproductivity, and human health. To better understand the dispersion and the impacts of Saharan dust as well as biomass aerosols of African origin, a set of ship-based research expeditions called the trans-Atlantic Aerosol and Ocean Science Expeditions (AEROSE) have been conducted in 2004 and annually since 2006. This report will discuss some of the findings of AEROSE-V, which took place during July 11 – August 10, 2009. Ambient aerosol mass densities were measured using a Quartz Crystal Microbalance Cascade Impactor (QCM). Size-segregated samples were also collected in order to study the compositional variability in the ambient mass distributions observed over the tropical Atlantic Ocean. In situ aerosol measurements were compared with outputs and analyses from NAAPS forecast models, satellite imagery, and other atmospheric data acquired aboard the ship. Preliminary analysis indicates that three distinctive air masses (biomass aerosol, Saharan dust, mixed zone) were encountered and distinguished by unique mass density distributions and evolution. The biomass aerosol appears to have originated from extensive wildfires burning in Central Africa.

Atia, et al. p.3

[.]1. INTRODUCTION

The Saharan Desert is the source of nearly three billion metric tons of mineral aerosols entering the atmosphere annually (Prospero et. Al, 1996). These dust aerosols, along with biomass aerosols originating from natural and human-induced processes in west and central Africa, and aerosols originating from urban and industrial activities, are carried by the Northeast Trade Winds across the Atlantic Ocean in to the Caribbean, South America, and eastern seaboard of the US. The impacts of these aerosols on our atmosphere, ocean and health manifest in regional impacts on climate and weather, deposition of harmful species in the ocean mixed layer, and through exposures to fungi, bacteria and air toxics that may induce and exacerbate asthma and other respiratory ailments.

While Saharan dust transport may be observed through satellite monitoring, in situ data of aerosol transport over the remote tropical Atlantic are scarce. The Class I NOAA research vessel, Ronald H. Brown, served as the scientific platform for the trans-Atlantic Aerosol and Ocean Science Expedition (AEROSE) campaigns. The general objectives of AEROSE are to characterize the chemical and microphysical properties of Saharan dust and biomass burning aerosols and investigate the impacts of these aerosols during trans-Atlantic transport. By conducting shipboard campaigns, essential in situ data was gathered, including data for satellite and sensor validation, another key goal to be achieved during these expeditions.

The fifth Aerosol and Ocean Sciences Expedition (AEROSE-V) was conducted from July 11 through August 10, 2009. The cruise began in Bridgetown, Barbados and concluded in Key West, Florida. The focus of this paper will be on size-resolved aerosol mass densities retrieved during the AEROSE-V campaign. Daily averages of these data were calculated throughout the cruise to help monitor air mass types and transitions during transect. The data collected from the west-east and south-north transects were quantified individually to observe aerosol concentration as a function of time and position. Combined with single-particle analysis of microphysical and chemical characteristics, this information can help improve forecast models for weather, climate, and air quality.

Figure 1. The AEROSE-V cruise began on July 11 at Bridgetown, Barbados and concluded at Key West, Florida on August 10. Each red marker consecutively represents the day as well as the approximate location of mass density measurements collected.

2. Measurement Strategy

During the first few days aboard the Ronald H. Brown vessel, instrumentation was setup and calibrated. The Quartz Crystal Microbalance (QCM) Cascade Impactors, California Measurements, were placed inside the Howard University van (located on the forward 02 level of the ship). The 02 level is about 8 meters above mean sea level, and the van is approximately 2.5 meters high. The vacuum line was connected from the intake valve of the QCM cascade impactor and extended outside to a PM2.5 impactor about one meter above the van. ………………………………

Two models of the QCM cascade impactor were used: the six-stage and ten-stage. Each impactor analyzes particle-size distributions and mass concentrations semi continuously at a flow rate of 2 lpm. Samples can be preserved for single-particle analysis. The QCM impactors were implemented based on the type of air mass transected. The six-stage impactor was typically used for dusty air masses due to the coarse-sized particles of dust, while the ten-stage impactor was used for biomass-burning air masses due to the finer particulate matter that is typical of smoke. The six-stage model consisted of the following size bins: 0.15, 0.3, 0.6, 1.2, 2.5, and 5.0 microns. The ten-stage model consisted of these size bins: 0.05, 0.1, 0.2, 0.3, 0.5, 1.0, 2.0, 3.0, 5.4, and 10.0 microns. These sizes represent the aerodynamic diameters of

particles within ±20% accuracy. ……………………………………………

Since the QCM cascade impactors take semi continuous measurements, samples were collected at one-minute, three-minute, and five-minute intervals depending on the predicted level of aerosol concentration in an air mass. In respect to mass density sampling by the QCM, NAAPS aerosol models were one of the main resources used for forecasting surface-level aerosol concentrations and air mass types (Saharan dust, biomass aerosol, or mixed ). Air masses with low concentrations were measured in long time-intervals and high concentrations were measured in short time-intervals because the instrument would output erroneous data once the sampling substrates reached saturation. Three runs were conducted and averaged for each measurement. Since our motive was to conduct analysis as a function of position and time, all measurements halted when the ship was stationary. When the smokestack of the ship moved towards the direction of the instrument, sampling was also stopped to avoid contamination of the data .

Atia, et al. p.3

Aerosol mass density sampling began on July 16 and concluded on August 2 (Figure 1). However, samples retrieved after July 30 were strictly stored for single particle analysis and were not quantified due to extensive sampling times.

3. Results and Discussion

Quantitative sampling was conducted from July 16-July 30 (Figure 2). Figure 2 shows mass density distributions as well as air mass types encountered throughout the duration of the cruise. On July 16, the ship crossed the ITCZ. However, from July 16-July17, there seemed to be a light mixed zone of aerosols. On July 18, the transition into a smoky air mass, which was believed to be the result of biomass burning, was observed and further indicated by the bimodal distribution of particle size concentrations. This air mass was transected until July 22. During this time period, a notable increase in concentration of the finer particulate matter was observed. From July 23-July 24, the ITCZ was crossed a second time. July 24 showed some interesting mass distributions including a large peak in the 2.5 micron size fraction. However, due to the complex conditions within the ITCZ, further analysis must be conducted to have a better comprehension of

the results from that day. There was then another air mass transition into a mixed aerosol zone on July 25 followed by a dust regime which was encountered and measured from July 26-July 30. Peaks generally occurred in the coarse-sized particle range, which was expected for dust. No measurements were taken by the QCM on July 28 because the ship was stationary for the majority of that day.

Table 2 shows daily averaged mass density values for each size fraction. July 21 and July 26 were the two peak days for smoke (biomass burning) and Saharan dust aerosols, respectively. On July 21, smoke peaks in the 0.2 micron size fraction with a concentration of 4.546 micrograms∙meter-3. On July 26, dust peaks in the 1.2 micron size fraction with a concentration of 18.107 micrograms∙meter-3. While these peaks in concentration were assumed to be indicative of smoke and dust, respectively, because of particle size, it is important to note that single particle analyses must be performed on these samples in order to verify the suspected aerosol classification.

The daily averaged PM 1.0 and 2.5 values quantified throughout the cruise are shown in Figure 3. A trimodal distribution is seen where

Atia, et al. p.3

Date / .05 um / .1 um / .15 um / .2 um / .3 um / .5 um / .6 um / 1.0 um / 1.2 um / 2.0 um / 2.5 um / 3.0 um / 5.0 um
7/16/2009 / 0.203 / 0.722 / 0.987 / 2.383 / 2.974 / 0.341
7/17/2009 / 0.223 / 1.528 / 1.417 / 3.149 / 3.346 / 0.354
7/18/2009 / 0.402 / 2.170 / 0.595 / 2.420 / 2.718 / 0.245
7/19/2009 / 0.255 / 1.202 / 0.468 / 1.826 / 2.061 / 0.282
7/20/2009 / 0.213 / 0.860 / 0.265 / 2.504 / 2.104 / 1.152 / 1.458 / 2.622 / 4.010 / 3.384 / 6.712 / 1.245
7/21/2009 / 1.059 / 0.316 / 4.546 / 2.983 / 1.784 / 1.736 / 3.775 / 4.513 / 3.326 / 4.465 / 0.833
7/22/2009 / 0.381 / 2.546 / 1.802 / 2.789 / 6.275 / 1.151
7/23/2009 / 0.388 / 2.475 / 2.202 / 3.401 / 6.343 / 0.465
7/24/2009 / 0.933 / 0.091 / 0.242 / 0.454 / 1.433 / 0.771 / 9.108 / 1.902 / 6.740 / 2.298 / 19.373 / 1.886 / 1.867
7/25/2009 / 0.798 / 0.971 / 1.630 / 4.748 / 3.041 / 0.943
7/26/2009 / 0.623 / 0.487 / 1.093 / 1.819 / 3.470 / 8.888 / 8.839 / 18.107 / 5.908 / 10.042 / 1.861 / 0.128
7/27/2009 / 2.869 / 0.829 / 1.367 / 5.372 / 10.114 / 9.812 / 2.155
7/28/2009
7/29/2009 / 0.652 / 2.359 / 5.309 / 8.859 / 8.473 / 1.721
7/30/2009 / 0.207 / 3.298 / 7.754 / 15.306 / 16.425 / 2.132

Atia, et al. p.3

Table 1. Daily averaged mass density values for each size fraction (micrograms∙meter -3 ).

Atia, et al. p.3

Figure 3. Daily averaged PM 2.5 and PM 1.0 values from July 16-July 30.

peaks occur on July 21, July 24, and July 26 in a smoke regime, the ITCZ, and a dust regime, respectively. Further analysis must be conducted to determine the aerosol type encountered during the transect of the ITCZ although a mixture of dust and biomass burning aerosols is suspected. Figure 4 shows the eastward transect of AEROSE-V from July16-July 20. Figure 5 shows the northward transect from July 20-July 27. These data represent hourly-averaged mass density values which were plotted as a function of position in the PM 1.0 and 2.5 size bins. Although the data points are connected by curves merely to show the trend of aerosol concentration as the ship gets closer to the source region, one should bear in mind that the QCM measurements were taken semi continuously. As the vessel approaches the west African coastal region, it is apparent that aerosol concentration increases, especially. as the vessel continues northward near Cape Verde, parallel to the west African coast.

NAAPS aerosol models were used to forecast aerosol regimes during the cruise. These models shared a fair correspondence with the data that was collected by the QCM. Highlighting the two days of peak smoke and dust concentrations, respectively, Figures 6 and 7 show the NAAPS aerosol predictions on July 21 and July 26. On July 21, the ship was transecting a smoke plume where the surface concentration was in the range of 2 to 8 micrograms∙meter -3 according to the NAAPS

model. On July 26, the model shows a range of 40 to 320 micrograms∙meter -3 for surface concentration of dust.

For the sake of comparison, if an ideal assumption is made concerning the QCM data, where smoke represents the particle size range of 0.05 to 0.5 microns (finer particulate matter), and dust represents the range of 0.6 to 3.0 microns, then July 21 yields a value of 10.689 micrograms∙meter -3 for the surface concentration of smoke, and July 26 has a value of 53.645 micrograms∙meter -3 for the surface concentration of dust. Considering this assumption, on these days, dust fits within the predicted range of NAAPS while smoke is slightly above the range of the model. A more accurate comparison may be made after the QCM samples undergo single particle analysis.

Atia, et al. p.3

Figure 8. Satellite image retrieved by Aqua MODIS on July 20, 2009 over the Kasai region of the Democratic Republic of Congo where red symbolizes areas where there are fires (Image by Jeff Schmaltz, MODIS Rapid Response Team).

While it was known that smoke and dust outflows were encountered during AEROSE-V, an effort was made to estimate possible source locations for these aerosols. On July 20, AQUA MODIS captured an image over the Kasai region of the Democratic Republic of Congo (Figure 8). This image unveiled the most probable source of the biomass burning aerosols, which seemed to be the result of numerous fires within this region and proximal regions within Angola, Zambia, and Tanzania. In fact, after observing NAAPS aerosol models focused over central and southern Africa throughout the month of July, smoke plumes seemed to constantly arise from this regional area. These fires are due to agricultural and forest burning during the African dry season. In addition to this evidence, a HYPSLIT model was used to observe the back trajectory plot of the smoky air mass. The model was set to calculate a back trajectory for a time duration of 72 hours. However, as Figure 10 shows, the trajectories end over the ocean and do not lead back to central Africa. Running the model for longer time durations would produce less accurate results. Therefore, wind data was implemented to complement the HYSPLIT back trajectory. Data sets collected by Meteosat-9 (Figure 11) and the SeaWinds Scatterometer (Figure 12) were used for this purpose. The grid in Figure 12 shows wind vectors at sea level from 10°S to 20° S and 0° to 15° W while the Meteosat-9 image shows low-level wind vectors within a wider spatial area. From these data, the net direction of southeasterly winds can be observed, and a fair assumption can be made that biomass burning aerosols were transported towards the measurement location by these winds.