December 31, 1999
The Asian Dust Events of April 1998
R. B. Husar1, D. M. Tratt2, B. A. Schichtel1, S. R. Falke1, F. Li1 D. Jaffe3, S. Gassó3, T. Gill4, N. S. Laulainen5, F. Lu6., M.C. Reheis7, Y. Chun8, D. Westphal9, B. N. Holben10, C. Geymard11, I. McKendry12, N. Kuring10, G. C. Feldman10, C. McClain10, R. J. Frouin13, J. Merrill14, D. DuBois15, F. Vignola16, T. Murayama17, S. Nickovic18, W. E. Wilson19, K. Sassen20, N. Sugimoto21
Abstract. In April 1998, several unusually intense dust storms were generated over the Gobi Desert by springtime cold weather systems with over 20 m/s surface wind speed. The dust cloud from the April 19 storm was swiftly transported across the Pacific reaching North America within 5 days. Part of the cloud subsided to the surface between British Columbia and California while another part was observed aloft in layers up to 10 km. During the peak on April 29 the excess dust aerosol concentration over the West Coast was about 20-50 mg/m3 and the aerosol concentration approached the health standard (daily PM10, 150 µg/m3). The dust chemical composition over North America was uniform and had a volume mean diameter of 2-3 mm. Throughout its atmospheric residence of two weeks, the dust cloud increased the surface albedo over the cloudless ocean and land. However, during a large part of its atmospheric residence the dust was evidently on top of low-lying clouds and significantly reduced the cloud albedo particularly near the UV. Over the West Coast, the dust scattering had redistributed the direct solar radiation into diffuse sky light. The Asian dust was detected by its distinctly yellow appearance on SeaWiFS and its evolution was monitored by satellites and surface monitoring, but these routine and serendipitous observations were incomplete. The dust event was observed and interpreted by an ad-hoc international web-based virtual community of scientist. This is an early summary of the dust event prepared by the virtual community.
Extreme biogeochemical events such as volcanic eruptions, forest fires, and dust storms provide unique opportunities to examine the inner workings of the atmospheric system. Such events tend to produce large quantities of dust, smoke, or haze, which is then dispersed over regional or global scales and serve as tracers for the events. The easily observable atmospheric particles can also visualize and quantify the nature of transport, transformation and removal processes along their path. This is particularly useful for establishing the link between local and global scale atmospheric processes. An early example for aerosols as visualizers was the discovery of the global circulation of the atmosphere. The organized global flow pattern was recognized through the globally observed stratospheric aerosols and red sunsets caused by eruption of the Krakatoa volcano in 1883 [Kiessling, 1885]. Clearly, the atmospheric particles from these events are not mere neutral tracers and visualizers, but are also major carriers in the biogeochemical cycle of sulfur, nitrogen, carbon and trace metals, as well as crustal elements. Last but not least, dust, smoke and haze particles produced by these processes change the climate and weather, fertilize the oceans and land, and cause health effects to humans.
The eruption of volcanoes, major forest fires and large dust storms are often associated with catastrophic consequences to humans and their environment, but the prediction of their specific occurrence in space and time is virtually impossible. Continuous monitoring of indicator signals and early warning systems has been effective in minimizing some of the risk to public health and safety. The unpredictability of these events also presents a unique challenge to atmospheric research, since sustained readiness for intensive field campaigns is almost impossible to maintain. Hence, the study of such events generally has to rely on the integration of routine monitoring data and serendipitous observations.
In April 1998, several major dust storms occurred over East Asia. In particular, the dust storm that occurred on April 19, 1998 over Mongolia and North-Central China produced a dust cloud that crossed the Pacific and caused aerosol concentration near the range of health standards (daily PM10 > 150 mg/m3) over much of the West Coast of North America. Fortunately, the dust events were detected and followed by an ad-hoc international group of investigators. This JGR issue contains a number of reports on various aspects of these dust events. This paper summarizes the major overall features of the April 98 dust events, with special emphasis on the Asian dust observations over North America.
Wind blown dust originating from the arid deserts of Mongolia and China is a well-known springtime meteorological phenomenon throughout East Asia. In fact, “yellow sand” meteorological conditions have acquired local names: Huangsha in China, Whangsa in Korea, and Kosa in Japan. The Asian dust storms have been studied for decades to understand their sources, mechanisms of transport, as well as the aerosol characteristics, including the effects on radiation. [Mizohata and Mamuro, 1978; Zhou et al., 1981; Wang et al., 1982; Iwasaka et al., 1983; Zaizen et al., 1995; Zheng et al., 1998; Zhang and Lu, 1999]. However, quantitative understanding of individual dust events e.g. the dust emission locations and rates is still incomplete.
The transport of desert dust from Asia to the North Pacific atmosphere is well documented [e.g. Shaw, 1980; Duce et al., 1980; Parrington et al., 1983; Uematsu et al., 1983; Merrill et al., 1989; Bodhaine, 1995; Husar et al., 1997] and show a maximum in dust loading each spring. Over the Pacific, the concentration of species from anthropogenic sources in Asia was also found to be enhanced during spring [Prospero and Savoie, 1989; Jaffe et al., 1997; Talbot et al., 1997] and have been documented to reach North America [Jaffe et al., 1999].
Compelling geological evidence of global scale transport of Asian dust emerged from the chemical analysis of samples in the Greenland ice core [ref] and Hawaiian soil studies [Rex et al., 1969; Dymond et al., 1974; Kennedy et al., 1998; Chadwick et al., 1999]. The elemental fingerprints of deposited dust at both locations were most consistent with the composition of the Asian dust sources.
Several keen observers in Asia and North America monitored the unusual dust cloud independently. In Japan for example, an integrated lidar monitoring network was operational in anticipation of the springtime Kosa season [Murayama et al., this issue]. At CAPITA the dust cloud was detected by routinely scanning the daily SeaWiFS color satellite images using the web-based NASA Goddard SeaWiFS Global Browser facility. The dust appeared as a distinct yellow cloud while it was passing over eastern China on April 16, 1998 (Figure 2b). Independently, the dust cloud was also detected over the western Pacific by the examination of GOES9 geostationary satellite images [Bachmeier, 1998].
When it was evident that the Asian dust cloud was reaching North America, an interactive website was set up at CAPITA, Washington University, (http://capita.wustl.edu/Asia-FarEast/) on April 26 to register various observations, exchange opinions and to support general communication and data sharing (cooperation) among the interested observers. By April 29, through word of mouth, over 40 scientists and air quality managers from North America and Asia had learned about the website, signed up to the ad-hoc workgroup, and began sharing their data and qualitative observations with the spontaneously formed virtual community. Most participants have maintained their data and preliminary reports on their own web pages but there was a user-maintained central catalog of the existing reports and discussions. A preliminary summary report was web-published by the ad-hoc workgroup within two weeks of the dust event, on May 11, 1998. For the following 18 months, individual research groups proceeded with their respective analysis without special coordination. The website was dormant but it served as the accessible repository of the dust event-related information. In December 1999, the virtual workgroup consisting of the authors of this paper was re-activated to produce this summary report.
The dust events were observed through routine satellite sensors, lidar instruments, sun photometers, airborne samplers, and a large array of surface-based aerosol monitors on both sides of the Pacific. Furthermore, dynamic dust transport models were used to simulate and to interpret the observations.
Remote sensing of the dust events. Satellite remote sensing was crucial in detecting the dust clouds, following their evolution as well as characterizing the spatial, temporal, and optical characteristics of the moving dust clouds. Routine quantitative techniques for the retrieval of aerosol properties over land from satellite data are not yet available, but it is an area of active research [King et al., 1999]. For this reason, most of the satellite data are used here as qualitative indices.
Spectral images of the dust cloud from the SeaWiFS sensor [McClain et al., 1998; Barnes et al., 1999] provided detailed spatial and spectral pattern of the dust at about local noon each day. On the SeaWiFS images the dust cloud was recognized by its bright yellow color, partial transparency and smooth spatial texture (Figure 2a-d). The raw (Level 1A) SeaWiFS data were obtained from the SeaWiFS Program [McClain et al., 1998] and processed at Washington University. The resulting spectral reflectance values (fraction of radiation reflected) represented the combined reflectance from the land, clouds and the aerosol. The scattering by air molecules was removed from the total reflectance using a procedure by [Vermote and Tanre, 1992], which also included nominal corrections for ozone and water vapor absorption.
The on-line, hourly, GMS-5 geostationary satellite data with visible and IR channels were used for near continuous monitoring of the dust plume dynamics throughout the day. Visible images from geostationary satellites (GOES8, GOES10, GMS-5) were most effective for dust detection during the twilight hours when the sun was in front of the sensor. At dawn or dusk, the reflectance from the ground surface is minimal while the reflectance from the aerosol is still high. This, along with forward scattering from the sun, tends to produce a strong amplification of the aerosol scattering signal (Figure 4).
During the post-analysis of the dust episode, operational retrievals of absorbing aerosol index from the TOMS satellite [Herman et al., 1997] provided useful information on the daily spatial distribution of the dust cloud. The TOMS aerosol absorbing signal is a semi-quantitative index of the columnar absorption by aerosols at 340 nm. The signal is derived from the absorption of the upwelling Raleigh scattering in the lower strata of the atmosphere. The April 98 dust clouds over Asia were also detected using the AVHRR sensor by the ‘split window’ method, taking the difference between the 11 and 12 mm channels [Murayama et al., this issue].
The vertical dust profiles were detected by sun photometers as part of the AERONET network in Asia and in North America [Holben et al., 1998], by the Solar Radiation Monitoring Laboratory, University of Oregon [Vignola et al., 2000] and by a number of other sun photometer networks.
The vertical dust profiles were monitored by lidar instruments on both sides of the Pacific. The passage of the April 15 dust cloud over Japan was monitored by a coordinated network of lidar instruments [Murayama et al., this issue] positioned throughout Japan, augmented by sun photometers and surface concentration measurements. The dust cloud was detected over North America as part of the routine long-term monitoring program of the aerosol vertical structure conducted by the CO2 backscatter lidar facility at the Jet Propulsion Laboratory [Tratt and Menzies, 1994]. Timely alerts provided by the ‘virtual community’ through the website allowed normal operations at this facility to be optimally focused on observation of the extreme dust event. Observations of the Asian dust cloud were also reported from lidar at University of Utah as part of their Cirrus monitoring program [Sassen, 1997].
In-situ measurements. The detailed physical, chemical and optical characteristics of the April 98 dust aerosol were recorded by surface-based in-situ measurements and aircraft samplings. Daily visibility observations, around the Gobi Desert provided an indication of reduced visibility due to the dust as well as the cause of the obstruction to vision, i.e. dust [NCDC, url, 1998].
Continuous aerosol size distribution data using an optical counter were reported from Korea [Chun et al., 2000]. More than 150 PM10 samplers were located throughout the west coast of North America. These samplers collected particulate mass less than 10 mm, some operating hourly, others collecting a 24-hour sample every sixth day [EPA AIRS]. The composition of the dust aerosol for chemical fingerprinting was captured in great detail by over fifty stations of the IMPROVE network that sampled twice a week, Wednesdays and Saturdays [Malm et al., 1994].
Modeling of the dust event. The April 1998 Asian dust event was simulated by at least four modeling groups. The dust models have provided valuable insights into the emissions, transport and removal process that governed the life cycle of such events [Westphal, 2000]. Also, the data from field observations were used to "test and validate" the performance of dynamic global scale aerosol models [Nickovic et al., this issue; Uno et al., this issue). The details of the dust subsidence over British Columbia were further investigated with the aid of a mesoscale transport model [McKendry et al., this issue].
The 1998 East Asian dust season, from March through May, was unusually intense. [Murayama et al., this issue]. In order to gain a perspective of the April dust pattern the daily time series of three aerosol measurements were examined: 1) TOMS satellite data averaged over the Gobi Desert and East Asia regions; 2) aerosol optical thickness measured at Dalamzagdad, Mongolia; and 3) surface extinction coefficient based on eight synoptic visibility monitoring stations in the Gobi Desert (Figure 1); and the mass of lifted and suspended dust particles with radii less than 5 microns as simulated by Westphal . The aerosol time series representing the Gobi Desert region convey consistently that the two largest dust storms of the season occurred on April 15 and 19. Furthermore, the TOMS data averaged over East Asia, indicate that the April 19, 1998 storm was the most intense dust event in the 1997-99 period. The simulations of suspended and lifted dust by Westphal  are generally consistent with the above dust emission pattern, with the April 19 event being the largest and longest lived event.
SeaWiFS satellite data indicate that the dust storms on April 15 and 19 originated from the same region of Gobi Desert (Figure 2a,d). On the satellite images the dust sources can be identified as streaks of dust plumes originating from specific patches of land, presumably from dunes of sand or loess not covered by vegetation. After about 500 km of transport, the plumes tend to merge and the streaky plume structure disappears. The location of the dust cloud is also seen in surface visibility observations reported by the global synoptic network, where on April 19, the visibility was reduced due to dust throughout central and eastern Mongolia (Figure 2d). The optical thickness of the suspended dust measured in Dalanzagdad, Mongolia [Holben, 1999] indicate that the turbidity increased from t < 0.5 on April 18 to t > 2 on April 19 as the dust cloud passed by.
Analysis of meteorological data and modeling results [Nickovic et al., this issue; Westphal, 2000] indicate that the formation of the April 15 and 19 dust clouds were associated with cold weather systems with high surface wind speeds. For example, as the April 19 storm swept through Mongolia and North-Central China, the surface wind speeds increased to over 20 m/s (Figure 2d). This was well above the generally used threshold wind speed (5-6 m/s) for dust suspension [Gillette, 1978]. The region of high wind speeds coincided with the Gobi Desert in Southern Mongolia and the Loess Plateau in the Gansu Province of China where the dust plumes were seen in the SeaWiFS data. As a result, by the time the storm reached eastern Mongolia, a 1,000 km long, optically thick dust cloud was generated with a sharp dust front (Figure 2d).
The most detailed spatial distribution of the dust is obtained from the 1 km resolution SeaWiFS satellite images. Figures 2a-c depict the dust pattern on April 15, 16, and 17 from the April 15 storm. Figures 2d-f show the corresponding 3-day images (April 19, 20, and 21) for the April 19 storm. For comparison, the magnitude of the TOMS absorbing aerosol index is superimposed as contours on the spectral SeaWiFS images. The spectral reflectance images from SeaWiFS provide a rich visual context relevant to the dust, including surface reflectance, position of cloud systems relative to the dust as well as indications of wind direction. A full interpretation of these images will have to await further analysis.
The dust clouds emitted on April 15 and April 19 had different meteorological driving forces and they followed distinctly different pathways. The April 15 storm was evidently caused by a “cut-off” low-pressure system, which is characterized by strong subsidence within the system, which tends to trap the dust near the surface [Murayama et al., this issue]. The dust cloud followed a southern route toward central China and subsequently turned toward Korea to the north. Throughout much of its residence over East Asia, the April 15 dust cloud was near or within precipitating cloud systems.
The April 19 dust storm was driven by a low-pressure cold front that entered western Mongolia and swiftly moved eastward, north of Korea, with the leading edge reaching the sea of Japan by April 20 [Westphal, 2000]. By the 21st, the leading edge of the dust cloud reached the Pacific. Furthermore, the yellow discoloration of the clouds indicates that at least part of the dust is located over the low-lying white clouds. Once the dust cloud reached the Pacific Ocean, it was carried by the westerly winds that are typical for the northern mid-latitudes (30-60 °N) in the springtime. The wavy dust transport path across the Pacific was visualized on the SeaWiFS images by the dust itself as a streaky yellow dye (Figure 3). It is evident that during the trans-Pacific passage, the dust cloud was stretched longitudinally. Also, large segments of the dust cloud were peeled off and transported northward into the Arctic reducing the amount of dust reaching North America. Model simulations [Westphal, 2000; Nickovic et al., this issue] also show such piece-wise disintegration of the dust cloud. The main dust cloud arrived at the West Coast of North America on April 25 and persisted until the beginning of May.
The visible channel of the GOES 10 geostationary satellite provides a unique view of the dust spatial distribution on the evening of April 27 at about 6 pm PST (Figure 4). Evidently, the dust cloud covered the entire west coast of North America from California to British Columbia and a wedge shaped region of the northwestern US and adjacent Canada. Near the US-Canadian border, streaks of dust indicate swift transport toward the upper Midwest and Ontario. (Figure 4). A second stream turned toward the south, blanketing the West Coast from British Columbia to California.
PM10 Monitoring data [EPA AIRS] over the U.S. West Coast provide a relatively detailed temporal pattern and a spatial map of the surface dust concentrations (Figure 5). Daily PM10 concentrations averaged over 150 AIRS stations throughout the west coast show a strong peak between April 26 and May 1 (Figure 5a). During the dust incursion, the regional average PM10 concentration reached 65 µg/m3 compared to 10-25 µg/m3 during the remainder of the April-May 1998 period. This suggests, that the excess dust concentration over the West Coast was about 40 µg/m3. On April 29, the contour maps show that the PM10 concentration over parts of Washington and Oregon exceeded 100 µg/m3 (Figure 5b). The low-lying areas of California, Nevada and Idaho also experienced concentrations well above 50 µg/m3. The dust concentration predicted by a long-range transport model compares favorably with the magnitude observed by the EPA AIRS PM10 network [Nickovic et al., this issue].
The spatial distribution of the Asian dust has also been established through the IMPROVE monitoring network data (Figure 7). Based on specific dust elemental signatures [Malm et al., 1994], the speciated aerosol data revealed the pattern of dust without interference from other local aerosol contributions, such as sulfates, organics, soot, and nitrates. On April 25, the western U.S. was virtually dust-free (Figure 7a), but reached high concentrations by April 29, particularly in the northwest (Figure 7b). On May 2, the elevated dust concentrations were highest over the Rocky Mountains and the Colorado Plateau (Figure 7c). By May 6, the dust levels had declined again throughout most of the western U.S. (Figure 7d). Photographs and observations from southeastern Utah taken on May 3-4 also record the progress of "strange haze" across the Colorado Plateau, moving from northern Arizona into western Colorado [Reheiss, url, 1998]. However, this dust transport to the interior of North America has not been confirmed by detailed meteorological analysis.
In southern British Columbia, spatial and temporal variations in PM10 concentrations were broadly consistent with those observed further south in the western U.S. PM10 levels increased dramatically to ~100 µg/m3 in the southern interior of the province on April 28. To the west, concentrations peaked in the Lower Fraser Valley (Vancouver region) on April 29 and then further west on Vancouver Island on April 30. This pattern is consistent with mesoscale modeling of the event [McKendry et al, this issue] showing strong subsidence and downward mixing of the dust layers over the mountainous interior and then westward surface transport of dust in "outflow" winds. Monitoring of aerosol light scattering at the Cheeka Peak Observatory in Washington state also showed the arrival of the dust cloud on April 28 with easterly flow. [Jaffe, 1999]. Based on analysis of elemental composition and meteorological analogues, it was estimated that Asian dust contributed 40-50% to peak observed PM10 levels in the Vancouver area [McKendry et al, this issue].
Hourly PM10 concentration data averaged over 12 stations in Northern California shows a strong and consistent diurnal cycle throughout the Asian dust episode between April 26 and May 2 (Figure 6a). Following the sharp concentration rise on April 26, there was a daily modulation with about 55 mg/m3 during the daytime hours (7AM-7PM) and a decline to 35 mg/m3 at night.
The April 1998 Asian dust event was a rare occurrence. Figure 8 shows the daily dust concentration at three IMPROVE monitoring sites for which long-term records exist, Mt. Rainier, WA, Crater Lake, OR, and Boundary Water, MN. At each site, the fine particle dust concentration was variable but generally below 5 mg/m3 throughout the decade of 1988-1998. The exception is the sharp dust peak at each of these sites reaching 25-30 mg/m3 on April 29, 1998. Hence, the April 1998 Asian dust event was more intense than any other dust incursion to the western U.S. over the past decade.
The limited data on the vertical dust transport and the resulting vertical profiles conveys an intricate pattern. The two major dust storms on April 15 and 19 appear to have started out with different vertical dispersion. A superposition of the TOMS and SeaWiFS data (Figures 2a and 2d) indicates significant differences during the first day of each dust storm. On April 19 the dust pattern from TOMS and SeaWiFS coincides geographically. On the other hand, the April 15 storm is clearly delineated on the SeaWiFS images, but the dust signal is absent from the TOMS UV-absorbing aerosol data. A full explanation for this difference is not available but it is conceivable that the dust cloud on April 15 was confined to the near surface layers, where the TOMS index is known to be less sensitive. One can speculate that absorbing aerosol of the dust front on April 19 was higher since it was fully detected by the TOMS sensor.
The complete vertical concentration profiles of the dust are not known, but lidar data for the April 15 event collected over Eastern China, Korea and Japan [Murayama et al., this issue] indicate a generally well-mixed dust layers from the surface to about 3 km. Unfortunately, the April 19 dust storm passed to the north of the East Asian lidar network, and lidar profiles are not available. By the time the dust cloud reached the East Coast of Asia, on April 21, some of the dust was elevated over the low-lying clouds as evidenced by the yellow coloration of white clouds. Similar qualitative observations over the Pacific also indicate that some of the dust was above the cloud layers but the actual dust heights over the Pacific are unknown.
A variety of dust observations over North America indicate a more layered structure. Serendipitous aircraft sampling of the arriving dust on April 27 near Seattle, WA show a distinct dust layer at about 2-3 km altitude [Gassó, 1999] and virtually no dust below. On the other hand, another aircraft sounding over eastern Washington State on May 1, indicates a surface-based dust layer up to about 2 km [Laulainen et al., 2000]. Aerosol Optical Depth (AOD) measurements at two elevations (1,088 and 100 m) are consistent with an aerosol scale height of about 2 km. Evidently, this second aircraft sampling on May 1 occurred after the dust layer had subsided.
A unique feature of the dust transport over the West Coast is the subsidence of the dust layer from the mid troposphere to the surface, between April 27 and 29 [McKendry et al., this issue]. In British Columbia, a zonally-oriented jet core to the north of Vancouver helped generate substantial mountain waves and strong subsidence (~0.05m/s) in the lee of the coast mountain range and the Rocky Mountains. Rapid downward transport permitted interception of dust layers by surface based mixing processes and then coastward transport by easterly surface winds. As a result of such processes, high surface level dust concentrations, lasting from April 27 until May 1, were observed over much of the northwestern U.S. and adjacent Canada.
Lidar data from Pasadena, CA and Salt Lake City, UT indicate dust layers at higher elevations. Figure 9a shows a superposition of the first (April 27) and final (May 1) lidar backscatter profiles observed from JPL in Pasadena, CA during the extreme dust event [Tratt et. al., this issue.]. On April 27, the dust layer was observed between 6 and 10 km. Note that the backscatter within the dust layer was about two orders of magnitude above the nominal prevailing background values. The University of Utah polarization lidar [Sassen, 1997] (Figure 9b) shows an aerosol layer at about 7.5 km altitude on April 24, 23:50 UTC at Salt Lake City, UT. The maximum depolarization of ~0.2 in the elevated aerosol layer is unusually high for this location and indicates the presence of super-micron sized non-spherical particles in the 1-2 mm diameter size range [Mishchenko and Sassen, 1998]. The elevated dust layer was easily visible from the ground, produced a solar aureole, and at times resembled very thin cirrus. Evidently, the dust layer detected at Salt lake City, UT was somewhat before the arrival of the main dust cloud on April 25.
The various dust aerosol observations facilitate the drawing of a limited number of inferences about the transformation and removal processes that affected the dust. The dust size distribution measurements in Korea [Chun et al., 2000] and aircraft sampling over Washington State [Gassó 1999; Laulainen et al., 2000] show that the aerosol volume distribution has a peak at about 2-3 mm (Figure 10). In Korea, the fraction of dust aerosol mass above 10 mm was insignificant and temporally uncorrelated with the dust 1-10 mm size range. This implies that the coarse particles above 10 mm, omni-present at the dust source, were preferentially removed by gravitational settling during the 3-7 day atmospheric transport time from Gobi to Korea and the West coast of North America.
The dust concentration in California (Figure 6) shows significant diurnal modulation, with reductions at night. Evidently, the dust reservoir remained in the boundary layer for six days and was well mixed to the surface during the daytime. However, at night data show regular concentration decreases. Possibly, the particles were removed at night by dry deposition within the shallow nocturnal layer without the replenishing mixing from above. [Wilson and Stockburger, 1990].
There were several observations of dust wet removal. The April 15 dust cloud appears to have been strongly depleted by wet removal processes throughout much of its residence over East Asia since it was embedded in a low pressure system [Murayama et al., this issue]. A yellow muddy rain with relatively high pH value (6.22) was reported on April 17 in Korea [Chun et al., 2000]. On April 17, the newspapers in Beijing, China also reported yellow muddy rain [Li, 1998]. However, detailed rain chemical composition data for the April dust events are not available at this time. One of the unique aspect of the April 19 dust cloud is that during the trans-Pacific transport there was apparently no significant wet removal by cloud scavenging and precipitation. This is inferred from the daily SeaWiFS images depicting the trans-Pacific transport (Figure 3).
The IMPROVE network data over North America show that the Asian dust chemical components are not correlated with particulate sulfur during the dust event. In fact, the samples with the highest dust concentrations have virtually no sulfur content. Aircraft sampling of the dust near Richland, WA on May 1, showed that the aerosol had no volatile components up to 300şC, which is consistent with a purely soil derived chemical composition of the dust [Laulainen et al., 2000]. Thus, based on the above crude estimates there is no evidence that the April 19 dust cloud had significant chemical transformations due to interaction with pollutant species, such as sulfur and nitrogen.
The properties of the April 1998 Asian dust were captured by in-situ and remote sensing instruments on both sides of the Pacific. The dust size distribution measured in Korea on April 19 and in Southern California an April 27 (Figure 10) evidently characterizes the aged dust emitted on April 15. In Korea, the volume size distribution function shows a sharp peak between 1-5 micron, with a volume diameter of 2 mm, and a logarithmic standard deviation of 1.6 [Chun et al., 2000]. In Southern California, the peak is shifted toward larger sizes and centered at 3 mm. Aircraft measurements through the dust layer near Richland, WA on May 1, exhibited a similar volume distribution [Laulainen et al., 2000]. Size segregated dust samples at numerous remote locations over the Northwest and adjacent Canada show that about 30-50% of the dust mass is below 2.5 mm [McKendry et al., this issue], which is consistent with the mass median diameter to 2-3 mm.
In Korea, continuous monitoring of particle concentration in different size ranges shows the strong correlation between the particles in the dust peak size range (2-3 mm) and virtually no correlation with particles below 0.8 mm and above 10 mm. Hence, the size ranges below 0.8 and above 10 mm have different origin than the coherent dust size range between 1-10 mm [Chun et al., 2000].
The chemical composition of the sampled dust over North America shows that the dust is composed of crustal elements with very consistent elemental ratios throughout the episode [McKendry et al., this issue]. Aerosol samples collected in aircraft over Washington State on May 1 were non-volatile and refractory in nature [Laulainen et al., 2000], which is consistent with crustal particle composition.
The spectral optical thickness of the dust measured in Dalanzagdad, Mongolia on April 19, was virtually constant between 0.4 and 1.02 mm (a ~ 0) [Holben, 1999], which indicates a characteristic size in excess of 2 mm. The dust optical thickness at Reno, NV clearly shows the arrival of the dust cloud. The aerosol optical thickness between April 25-29 was = 0.3 < t <0.5 at 525 nm, compared to 0.05-0.1 on the preceding days [Du Bois, 2000]. The spectral optical depth also increased at San Nicolas Island in Southern California during April 25-28, with values reaching 0.5 at 500 nm. Values of 0.4-0.5 were measured over the Pacific Northwest during the same period. For both locations the data show a small wavelength dependence of the optical depth, but the dependence was more pronounced at San Nicolas Island (alpha of -0.5 to -0.3) than in the Pacific Northwest (alpha of -0.3 to -0.2). Tratt et al. [this issue] report coarse mode radius retrievals for the dust of 1-2 µm based on sun and sky radiance measurements at San Nicolas Island. In the Pacific Northwest, the smaller wavelength dependence implies larger particle sizes, probably in excess of a few micrometers [Laulainen et al., 1999].
A distinct feature of the observed dust is the distinct yellow color. The color of optically thin aerosol clouds arises from light scattering and absorption of solar radiation and depends strongly on the scattering angle q (defined by the sun, dust cloud and the sensor). For the SeaWiFS satellite data the angle ranges between 130 < q < 170. The dust color is best detected over the ocean where the spectral albedo of the surface is low. Figure 11, shows the excess dust reflectance, i.e. the measured total reflectance with the ocean reflectance and Raleigh scattering subtracted. The data for the China Sea, Western Pacific and S. California indicate that near Asia the albedo of the dust cloud is in the range 0.1 – 0.2 but there is a sharp drop in the dust reflectance, particularly below 0.5 mm wavelength. In fact, even a thick dust cloud backscatters only about 5% of the incoming solar radiation. This drop in reflectance cannot be explained by purely Mie scattering, but it is not clear what are the relative roles of particle absorption in the near UV and irregular particle shape.
The most notable direct effect of the dust on the earth-atmosphere system is the increase of the surface albedo due to the backscattered radiation. The increased reflectance effect is illustrated using the 8 wavelengths (0.412-0.87 um) of the SeaWiFS satellite. The data show that dust-induced increase in the spectral albedo is shifted toward the red both over the ocean and land. Over arid land (Figure 2d) the dust cloud has increased the surface albedo in the red from 0.25 to 0.4-0.5 and in the blue from 0.1 to 0.3. Over the cloud-free ocean, the dust cloud has increased the apparent surface albedo from virtually zero to 0.1-0.2. However, there were also frequently layers of dust observed above the low-lying white clouds in which case the albedo of white clouds was reduced from 0.7 to 0.4-0.5, with preferential decrease in the blue. The often-observed yellow coloration of low-lying white clouds due to elevated dust layers eludes a plausible explanation but it served well for dust visualization. Since the dust absorbs part of the visible light, as evidenced by TOMS Absorbing Aerosol Sensor, the dust layer was heating the surrounding atmosphere during the sunny daytime hours.
During the five-day dust event in Oregon, there was a 25-35% decrease in normal solar radiation (Figure 12), although most of the loss from the normal solar beam was forward scattered radiation and reached the ground as diffuse radiation [Vignola et al., 2000]. As a consequence, the global broadband radiation onto a horizontal surface was reduced by the dust by less then 2% allowing < 2% for backscattering. This is consistent with SeaWiFS reflectance data, which indicate that the increase of the surface albedo near Eugene, OR from April 21 to April 27 was about 2%, or less.
The Asian dust storms of April 98 had major consequences to human health and welfare. According to CNN, on April 15, twelve people perished in the dust storm in the Xinjiang Autonomous Region of China alone. It is likely that this natural disaster has caused additional casualties along its path. Yellow muddy rain, reported over eastern China and Korea, produced an economic loss of unknown magnitude.
When the dust cloud reached British Columbia, Washington, Idaho and Oregon, the state health agencies issued air pollution advisory warnings to the general public with a corresponding ban on open prescribed burning. In making their decisions and informing the public, the state air pollution regulatory agencies have actively participated in the virtual community by supplying and using information on the virtual community website.
Such dust events have both research and regulatory implications. With a measured mass mean diameter of 2-3 um, the dust will influence PM2.5 as well as PM10 concentrations. This is important since there is increasing information to suggest that, at least for some health effects categories, fine-mode particles (<1 µm) are more likely to cause acute health effects than coarse-mode (>1µm) soil particles [Schwartz et. al., 1997, 1999]. Therefore, episodic intrusions of soil dust need to be identified not only to allow waivers for possible exceedances of the PM2.5 and PM10 standards but to correct PM2.5 time series for use in correlation of fine-mode particles with health effects.
The routine and serendipitous dust observations used in this summary (e.g., mass loading, particle size and chemical composition, aerosol optical thickness, dust layer height) have been brought to bear on the characterization of the dust in this particular event. However, a firm quantitative characterization of this event, particularly the physico-chemical processes, was not possible since many of the necessary measurements were not made. Also, it is not clear how representative the April 1998 events were. The planned ACE-Asia studies (http://saga.pmel.noaa.gov/aceasia) beginning in 2000 are expected to provide a rich characterization of east-Asian aerosol properties over time and above the surface, to better quantify the transport, transformations and fate and of these aerosols.
Although, the available observations are incomplete, much can still be learned through further data analysis and interpretation. Future activities may include collecting and organizing the available data into a documented and shared resource base. The collected observations could then be used for further model development, validation, and testing; evaluation of aerosol data derived from satellite observations; and other collaborative analysis.
The experience from this event could also help in planning more effectively disaster mitigation efforts, such as an early detection, warning and analysis system. The Asian dust event has demonstrated that the currently available space borne and surface aerosol monitoring can enable virtual communities of scientists and managers to detect and follow such major aerosol events, whether resulting from fires, volcanoes, pollution or dust storms. It has also been shown that ad-hoc collaboration of scientists is a practical way to share observations and to collectively generate the explanatory knowledge about such unpredictable events. A more robust and focused infrastructure that would facilitate such collaboration would assure that learning more about these unusual atmospheric events is not left to chance.
Bachmeier, S., Asian Dust Over the Pacific Ocean, April 22-24, 1998. http://cimss.ssec.wisc.edu/goes/misc/980424.html, 1998.
Barnes, R. A., R. E. Eplee, Jr., F. S. Patt, and C. R. McClain, Changes in the radiometric sensitivity of SeaWiFS, Appl. Opt., 38(21), 4649-4664, 1999.
Biscaye, P. E., F. E. Grousset, M. Revel, S. VanderGaast, G. A. Zielinski, A. Vaars, and G. Kukla, Asian provenance of glacial dust (stage2) in the Greenland Ice Sheet Project 2 Ice Core, Summit, Greenland, J. Geophys. Res., 102, C12, 26,765-26,781, 1997.
Bodhaine, B. A., Aerosol absorption measurements at Barrow, Mauna Loa and the South Pole, J. Geophys. Res., 100, 8,967-8,975, 1995.
Chadwick, O. A., L. A. Derry, P. M. Vitousek, B. J. Huebert, and L. O. Hedin, Changing sources of nutrients during four million years of ecosystem development, Nature, 397, 491-497, 1999.
Chun, Y., J., J. Kim, J. C. Choi, K. O. Boo, M. Lee, and S. N. Oh, Characteristic number size distribution of aerosol during Asian dust episode in Korea, submitted for publication, 2000.
Chun, Y. S. The physico-chemical features of dust particles collected in Korea during yellow sand episode, J. Geophys. Res., this issue.
Du Bois, D., Ph.D. dissertation, University of Nevada, Reno, 2000.
Dymond, J., P. E. Biscaye, and R. W. Rex, Eolian origin of mica in Hawaiian soils, Geol. Soc. Amer. Bull., 85, 37-40, 1974.
Duce, R. A., C. K. Unni, B. J. Ray, J. M. Prospero, and J. T. Merrill, Long-range atmospheric transport of soil dust from Asia to the tropical North Pacific: Temporal variability, Science, 209, 1,522-1,524, 1980.
EPA Aerometric Information Retrieval System (AIRS).
Gassó, S., personal communication, 1999.
Gillette, D., A wind tunnel simulation of the erosion of soil: Effect of soil texture, sand-blasting, wind speed, and soil consolidation on the dust production, Atmos. Environ., 12, 1,735-1,743, 1978.
Vignola, F. E., C. Gueymard, N. S. Laulainen, and J. K. Vaughan, Evaluation of the April 1998 Asian dust cloud with broadband and Narrowband irradiance methodologies, in preparation, 2000.
Herman, J. R., P. K. Bhartia, O. Torres, C. Seftor, and E. Celarier, Global distribution of UV-absorbing aerosols from Nimbus 7/TOMS data. J. Geophys. Res., 102, 16,911-16,922, 1997.
Holben, B. N., personal communication, 1999.
Holben, B N., T. F. Eck, I. Slutsker, D. Tanre, J.P. Buis, A. Setzer, E. Vermote, J. A. Reagan, Y. J. Kaufman, T. Nakajima, F. Lavenu, I. Jankowiak, and A. Smirnov, AERONET - A federated instrument network and data archive for aerosol characterization, Remote Sensing of Environment, 66, 1-16, 1998.
Husar, R. B., J. M. Prospero, and L. L. Stowe, Characterization of tropospheric aerosols over the oceans with the NOAA Advanced Very High Resolution Radiometer optical thickness operational product. J. Geophys. Res. 102, 16,889-16,909, 1997.
Iwasaka, Y., H. Minoura and K. Nagaya, The transport and spatial scale of Asian dust-storm clouds: A case study of the dust-storm event of April 1979. Tellus, 35B, 189-196, 1983.
Jaffe, D. A., A. Mahura, J. Kelley, J. Atkins, P. C. Novelli and J. Merrill, Impact of Asian emissions on the remote North Pacific atmosphere: Interpretation of CO data from Shemya, Guam, Midway and Mauna Loa. J.Geophys.Res. 102, 28,627-28,636, 1997.
Jaffe, D. A., T. Anderson, D. Covert, R. Kotchenruther, B. Trost, J. Danielson, W. Simpson, T. Berntsen, S. Karlsdottir, D. Blake, J. Harris, G. Carmichael, and I. Uno, Transport of Asian air pollution to North America, Geophys.Res.Letts. 26, 711-714, 1999.
Jaffe, D. A., personal communication, 1999.
Kennedy, M. J., O. A. Chadwick, P. M. Vitousek, L. A. Derry, and D. M. Hendricks, Changing sources of base cations during ecosystem development, Hawaiian Islands, Geology, 26, 1,015-1,018, 1998.
Kiessling, J., Die Daemmerungserscheinungen im Jahre 1883 und ihre physikalische Erklaerung, 64pp., Voss, 1885.
King, M. D., Y. J. Kaufman, D. Tanre, and T. Nakajima, Remote sensing of tropospheric aerosols from space: Past, present and future. Bull. Amer. Meteor. Soc., 80, 12,229-12,259, 1999.
Kuring, N. SeaWiFS observes transport of Asian Dust over the Pacific Ocean. http://daac.gsfc.nasa.gov/CAMPAIGN_DOCS/OCDST/asian_dust.html, 1998.
Li, F. personal communication, 1998.
Laulainen, N. S., J. K. Vaughan, F. E. Vignola, and R. J. Weber, Some physical and optical characteristics of the April 1998 Asian dust cloud over the Pacific Northwest, in preparation, 2000.
Malm, W.C., J. F. Sisler, D. Huffman, R.A. Eldred, and T. A Cahill, Spatial and seasonal trends in particle concentration and optical extinction in the US, J. Geophys. Res., 99, 1,347-1,370, 1994.
McKendry, I. G., J. P. Hacker, S. Sakiyama, D. Mignacca, and K. Reid, Transport of Asian dust to the Lower Fraser Valley, B.C., J. Geophys. Res., this issue.
McClain, C. R., M. L. Cleave, G. C. Feldman, W. W. Gregg, S. B. Hooker, and N. Kuring, Science quality SeaWiFS data for global biospheric research, Sea Technol., 39, 10-16 1998.
Merrill, J. T., M. Uematsu, and R. Bleck, Meteorological analysis of long range transport of mineral aerosols over the North Pacific, J. Geophys. Res., 94, 8,584-8,598, 1989.
Mishchenko, M. I., and K. Sassen, Depolarization of lidar returns by small ice crystals: An application to contrails. Geophys. Res. Lett., 25, 309-312, 1998.
Mizohata, A. and T. Mamuro, Some information about loess aerosol over Japan. Japan. Soc. Air Pollut., 13, 289-297, 1978.
Murayama, T., N. Sugimoto, I. Uno, K. Kinoshita , K. Aoki, N. Hagiwara, Z. Liu, I. Matsui, T. Sakai, T. Shibata, K. Arao, B.-J. Sohn, J.-G. Won, S.-C. Yoon, T. Li, J. Zhou, H. Hu, M. Abo, K. Iokibe, R. Koga, and Y. Iwasaka, Ground-based network observation of Asian dust events of April 1998 in East Asia 1. Transport mechanism, J. Geophys. Res., this issue.
Murayama, T., N. Sugimoto, I. Matsui, K. Arao, K. Iokibe, R. Koga, T. Sakai, Y. Kubota, Y. Saito, M. Abo, N. Hagiwara, H. Kuze, N. Kaneyasu, R. Imasu, K. Asai, and K, Aoki, Lidar network observation of Asian dust in Japan, Proc. SPIE, 3504, 8-15, 1998.
NCDC, Global Surface Summary of Day, National Climatic Data Center (NCDC) (http://www.ncdc.noaa.gov/), 1998
Nickovic, S. et al., Model for prediction of desert dust cycle in the atmosphere, J. Geophys. Res., this issue.
Parrington, J. R, W. H. Zoller, and N. K. Aras, Asian dust: Seasonal transport to the Hawaiian Islands, Science, 220, 195-197, 1983.
Prospero, J. M. and D. L. Savoie, Effect of continental sources of nitrate concentrations over the Pacific Ocean, Nature, 339, 687-689, 1989.
Reheiss, M., Possible Sighting of Asian Dust in Eastern Utah. http://climchange.cr.usgs.gov/info/dust/gobi/, 1998.
Rex, R. W., J. K. Syers, M. L. Jackson, and R. N. Clayton, Eolian origin of quartz in soils of Hawaiian Islands and in Pacific pelagic sediments, Science, 163, 277-291, 1969.
Sassen, K., Contrail-cirrus and their potential for regional climate change, Bull. Amer. Meteor. Soc., 78, 1,885-1,903, 1997.
Schwartz, J., D. W. Dockery, and L. M. Neas, Is daily mortality associated specifically with fine particles? J. Air Waste Manage. Assoc. 46, 927-939, 1997.
Schwartz, J., G. Norris, T. Larson, L. Sheppard, C. Claiborne, and J. Koenig, Episodes of high coarse particle concentrations are not associated with increased mortality, Environ. Health Perspect., 107, 339-342, 1999.
Shaw, G.E., Transport of Asian desert aerosol to the Hawaiian Islands, J. Appl. Met., 19, 1,254-1,259, 1980.
Talbot, R. W, J. E. Dibb, B. L. Lefer, J. D. Bradshaw, S. T. Sandholm, D. R. Blake, N. J. Blake, G. W. Sachse, J. E. Collins, B. G. Heikes, J. T. Merrill, G. L. Gregory, B. E. Anderson, H. B. Singh, D. C. Thornton, A. R. Bandy, and R. F. Pueschel, Chemical characteristics of continental outflow from Asia to the troposphere over the western Pacific Ocean during February-March 1994: Results from PEM-West B. J. Geophys. Res. 102, 28,255-28,274, 1997.
Tratt, D. M., and R. T. Menzies, Recent climatological trends in atmospheric aerosol backscatter derived from the Jet Propulsion Laboratory multiyear backscatter profile database. Appl. Opt., 33(3), 424-430, 1994.
Tratt, D. M., R. J. Frouin, and D. Westphal, The April, 1998 extreme Asian dust event: a Southern California perspective, J. Geophys. Res., this issue.
Uematsu, M., R. A. Duce, J. M. Prospero, L. Chen, J. T. Merrill. and R. L. McDonald, Transport of mineral aerosol from Asia over the North Pacific Ocean. J. Geophys. Res. 88, 5,343-5,352, 1983.
Uno, K., et al., Trans-Pacific dust transport, , J. Geophys. Res., this issue.
Vermote, E. and Tanre, D., Analytical expression for radiative properties of planar Raleigh scattering media, including polarization. J. Quant. Spectrosc. Radiat. Transfer, 47(4), 305-314, 1992.
Wang, M. X., J. W. Winchester, T. A. Cahill and L. X. Ren, Chemical elemental composition of windblown dust, 19 April 1980. Beijing. Kexue Tongbao, 27, 1193-1198, 1982.
Westphal, D. L., Real-time applications of a global multi-component aerosol model. Submitted for publication in J. Geophys. Res., 2000.
Wilson W.E. and L. Stockburger, Diurnal variations in aerosol composition and concentration in aerosols: Science, industry, health and environment, in Proceedings of the Third International Aerosol Conference, Kyoto, Japan, September 1990, S. Masuda and K. Takahashi, (Eds.), 962-965 pp., Pergamon Press,1990.
Zaizen, Y., M. Ikegami, K. Okada and Y. Makino, Aerosol concentration observed at Zhangye in China, J. Met. Soc. Japan, 73, 891-897, 1995.
Zhou, M., Q. Shaohou, S. Ximing and L. Yuying, Properties of the aerosols during a dust storm over Beijing area, Acta Scientie Circumstantae, 1, 207-219, 1981.
Zheng X., F. Lu, X. Fang, Y. Wang, L. Guo, A study of dust storms in China using satellite data" in optical Remote Sensing of the Atmosphere and Clouds, J. Wang, B. Wu, T. Ogawa, Z. Guan, Editors, Proc. SPIE, 3,501,163-168, 1998.
Zhang, D. and F. Lu, Winter sandstorm events in extensive northern China, (in Chinese), Quaternary Sciences, 5, 441-447, 1999.
18 Euro Mediterranean Centre on Insular Coastal Dynamics (ICoD), University of Malta, Malta.
20 Department of Meteorology, University of Utah, Salt Lake City, UT