Outbreaks of Asian Dust Storms An Overview from Satellite and Surface Perspectives

Si-Chee Tsay*

Laboratory for Atmospheres, Goddard Space Flight Center, Greenbelt, Maryland, USA

[email protected]

Taiwan, located downwind of dust storm outbreaks from China, in the sink region of biomass-burning aerosols from Southeast Asia, and at the outflow of urban-industrial pollutants from the Pearl and Yangtze River Delta, is exposed to a seasonal milieu of natural and anthropogenic aerosols in the atmosphere. In the springtime, outbreaks of Asian dust storms occur frequently in the arid and semiarid areas of northwestern China — about 1.6 x 106 square kilometers, including the Gobi and Taklimakan deserts — with continuous expansion of spatial coverage. These airborne dust particles, originating in desert areas far from polluted regions, interact with anthropogenic sulfate and soot aerosols emitted from Chinese megacities during their transport over the mainland. Adding the intricate effects of clouds and marine aerosols, dust particles reaching the marine environment can have drastically different properties than those from their sources.

Together with anthropogenic pollutants, airborne dust particles may alter regional hydrological cycles by aerosol direct/indirect radiative forcing, influence fisheries by causing nutrient deposition anomalies, and increase adverse health effects on humans by trace metal enrichment. In addition to their local-to-regional impact, these dust aerosols can be transported swiftly across the Pacific Ocean to reach North America in less than a week, resulting in an even larger scale effect. Asian dust aerosols can be distinctly detected by their colored appearance on modern Earth-observing satellites [e.g. MODerate-resolution Imaging Spectroradiometer, (MODIS), Total Ozone Mapping Spectrometer (TOMS), Sea-viewing Wide Field-of-view Sensor (SeaWiFS), Geostationary Meteorological Satellites (GMSs)], and their evolution monitored by satellites and surface networks [e.g. AErosol RObotic NETwork (AERONET), Micro-Pulse Lidar Network (MPLNET)]. However, these essential observations are incomplete due to the unique properties of the data constituted unilaterally on either the spatial (snapshot global coverage) or temporal (long-term point sites) dimension. Comprehensive modeling is required to bridge these spatial and temporal observations, and to serve as an integrator for our understanding of the effects of dust's physical, optical, and radiative properties on various forcing, response, and feedback processes occurring in the Earth-atmosphere system.

Recently, many field experiments (e.g. international ACE-Asia and regional follow-on campaigns) have been conducted to shed light on characterizing the compelling variability of dust aerosols on spatial and temporal scales, especially near the source and downwind regions. As a result of synergizing satellite, aircraft, and surface observations, our understanding of the distributions and properties of airborne dust aerosols has advanced significantly. It is our goal/hope to continue combining observational and theoretical studies to investigate in depth the changes of regional climate, hydrological budget, tropospheric chemistry, wind erosion, and dust properties in Asia. These regional changes (e.g. aerosol loading, cloud amount, precipitation rate) constitute a vital part of global change, and our success or failure in developing reliable predictions of, as well as adequate responses to, the changes will determine the prospective course for sustainable civilization. Consequently, the lessons learned will help strengthen our ability to issue early warnings of Asian dust storms and minimize further desertification in the future.

*With contributions from Gin-Rong Liu, Nai-Yung Hsu, Wen-Yih Sun, Neng-Huei Lin, Qiang Ji, Guang-Yu Shi, Myeong-Jae Jeong, Tang-Huang Lin, Chi-Ming Peng, Sheng-Hsiang Wang, and Jr-Shiuan Yang.

1. Background

Having entered the new millennium, there is no doubt that globalization is the wave of the future in societal and economic activities, including the byproducts associated with these activities. An excellent example is the recent intensified outbreaks of Asian dust storms and air pollution that have brought even broader public attention to their wide-ranging impacts, from the macro scale of aerosol radiative forcing on weather and climate to the micro scale of pathogens/minerals on the spread of human/agricultural diseases. Thus, aerosol emissions in one area can cause damage to other regions through transborder and even transcontinental transport. The stunning image of Living Earth (the Earthrise, first captured by NASA's Apollo-8 mission on 22 December 1968) rising above the Moon's horizon truly indicates the dynamical and restless nature of our home planet. Starting from the last century, however, the growth of the world population, the development of modern technology, the demand for consumable energy, and drastic land-cover/land-use changes, among other factors, have revealed an exponential trend that presents a major environmental stress that tips Earth away from sustainable balance. This urgent call for the saving of our planet has been echoed overwhelmingly by the recent documentary film An Inconvenient Truth, presented by Al Gore, former US vice-president and 2007 Nobel Peace laureate.

Taiwan, a role model for the economic miracle of developing countries during the mid-20th century, now faces many critical challenges regarding environmental protection. For instance, along with a multitude of other effects, the outbreaks of Asian dust storms seasonally alter the chemistry in the atmosphere and the surrounding seas, pollute the breathing air (impacting human health), and degrade the visibility (affecting traffic safety). In this overview article, we first give an introduction to the outbreak of Asian dust storms with historical and statistical prospects, as well as the environmental conditions associated with the dust source, transport, and sink. The societal impact and scientific significance of dust aerosols are presented in Sec. 2. Section 3 discusses the properties of dust aerosols based on available remote sensing measurements, in situ observations, and theoretical modeling to gain a better understanding of what role dust aerosols play in the Earth-atmosphere system. Finally, concluding remarks are given on what lessons have been learned and what further action strategies may be taken.

1.1. Dust storm outbreaks

Utilizing advanced instrumentation from space, large-scale dust storm outbreaks are commonly observed on the Earth and other planets in our solar system. For example, on 26 June 2001, NASA's Hubble Space Telescope spotted an enormous dust storm outbreak on Mars that quickly enveloped the whole planet and subsequently raised the temperature of the glacial Martian atmosphere by about 30° C. Although the sum of Earth's continents is roughly equal to the total Martian surface area, dust storm outbreaks on Earth are generally smaller because it is not a global desert like Mars. Nearly one-third of Earth's lands are deserts or arid regions, which receive annual precipitation of either less than 25 cm (—10 inches) or less than half of the evaporation. Based on this definition, the polar regions (—2.8 x 107 km2) comprise Earth's largest desert, followed by the Saharan (—9.1 x 106km2) and Arabian (—2.6 x 106km2) deserts in subtropical Africa/Mideast, and then China's Gobi-Taklimakan deserts (—1.6 x 106 km2) in midlatitude Asia.

A dust storm (or sandstorm in some contexts) outbreak has many names across the desert regions, such as "simoom" for the African Sahara, "haboob" for deserts in the Arabian Peninsula, " shachenbao" for the Chinese Gobi-Taklimakan deserts, and "kosa" for downwind

Figure 1. (a) Official document recording on bamboo strips in ancient China; (b) reconstructed image of a fierce dust storm originating at the Taklimakan desert that destroyed a stagecoach during the Chinese Han Dynasty, reported by a surviving officer (Xinhua News, May 2002); (c) a typical Taklimakan dust storm appearing in pale beige and sweeping toward the northeast, captured by MODIS/Aqua (image courtesy: NASA Earth Observatory on 3 December 2005); and (d) a few days after violent dust storm outbreaks from the Gobi-Taklimakan deserts, inhalable dust aerosols blanket the downwind regions of the Korean peninsula, Japan, and Taiwan (photo courtesy: Taipei Times on 20 March 2006 in Taiwan, with the Taipei-101 Tower, currently the world's tallest building, in the background).

Figure 1. (a) Official document recording on bamboo strips in ancient China; (b) reconstructed image of a fierce dust storm originating at the Taklimakan desert that destroyed a stagecoach during the Chinese Han Dynasty, reported by a surviving officer (Xinhua News, May 2002); (c) a typical Taklimakan dust storm appearing in pale beige and sweeping toward the northeast, captured by MODIS/Aqua (image courtesy: NASA Earth Observatory on 3 December 2005); and (d) a few days after violent dust storm outbreaks from the Gobi-Taklimakan deserts, inhalable dust aerosols blanket the downwind regions of the Korean peninsula, Japan, and Taiwan (photo courtesy: Taipei Times on 20 March 2006 in Taiwan, with the Taipei-101 Tower, currently the world's tallest building, in the background).

Korea-Japan regions. The prime mechanism causing these dust storm outbreaks differs noticeably: those from the Sahara are mainly driven by instability induced by strong solar heating at the subtropical surface (e.g. Karyampudi et al., 1999), with year-round frequency except during the African monsoon season; those forming at the Arabian deserts arise chiefly from the collapse of thunderstorms (also known as a downburst) during subtropical summer (e.g. Tindale and Pease, 1999); and those originating at the Gobi-Taklimakan deserts are largely associated with cold air outbreaks causing the Mongolian cyclonic depression and frontal activities in the spring (e.g. Qian et al., 2002).

As much as one-third to one-half of global dust emission, estimated to be about 800 Tg (Zhang et al., 1997), is introduced annually into Earth's atmosphere from various deserts in China. Asian dust storm outbreaks are believed to have persisted for hundreds of thousands of years over the vast territory of north and northwest China, but it was not until recent decades that many studies for compelling evidence for recognizing the importance of these eolian dust particles for forming the Chinese Loess Plateau (e.g. Derbyshire et al., 1998), and biogeochemical cycling in the North Pacific Ocean (e.g. Husar et al., 1997), to as far as in the Greenland ice-sheets (e.g. Svensson et al., 2000) through long-range transport. Recently, in the vicinity of Dunhuang, China — the gateway to the ancient Silk Road — Chinese archaeologists unearthed a "bamboo document" from the early Han Dynasty [see Fig. 1(a)], which described officially and explicitly the earliest dust storm event in Chinese written records near the source regions.

The Taklimakan ("place of no return" in Turkish) desert, a predominant land feature in the Tarim Basin, is enclosed by three major mountain ranges in western China: the Tien-Shan to the north, the Pamirs to the west, and the Kunlun-Shan to the south, with a narrow opening of saline marshy depression (the lowest area in the region is 150 m below sea level) in the east [see Fig. 1(c) for the topography]. Snow and glacier melt-waters from the surrounding mountain ranges supply all the rivers in the Tarim Basin, but these waters never find the sea. Added to this extraordinary geography, the location of the Taklimakan desert far from oceans further removes any rainfall from Asian monsoons. Having an annual precipitation rate of less than 10 mm and plenty of dry river/lake sediments, Taklimakan (^0.34 x 106 km2) constitutes the second-largest shifting-sand desert on Earth (e.g. Sun and Liu, 2006). As a result, it is hardly surprising that the Tarim Basin experiences more dust storms than any other place on Earth, with up to 100-174 events annually (e.g. Washington et al., 2003), as depicted in Fig. 1(c) for a typical dust storm outbreak. Another region of frequent dust storm outbreaks in China is Asia's largest desert area, the great Gobi ("very large and dry" in Mongolian, or simply "big desert" in Manchu), which is bounded by the Altai-Shan and Mongolia grasslands/steppes to the north, the Tibetan Plateau to the southwest, and the North China Plain to the southeast. Unlike the unique geography of the Taklimakan desert, the great Gobi (~1.29 x 106km2), by and large consisting of gravel and bare rock, nonetheless has numerous distinct ecological and geographic regions due to variations in local climate and topography. Outbreaks of dust storms in north and northwest China not only cover massive areas in source regions, but also transport airborne dust particles downwind to the Korean peninsula, Japan, Taiwan [see Fig. 1(d)], and further beyond (e.g. VanCuren and Cahill, 2002).

The spatial and temporal distributions of dust storm outbreaks in China have been the subject of numerous studies (e.g. Wang et al., 2004, and references therein). Generally, meteorological observers report a dust storm event when the horizontal surface visibility is reduced to 1 km or below, a blowing dust event for surface visibility in the range of 1-10 km, and a hazy sky for visibility less than 10 km with aeolian dust particles suspended homogeneously in the air. As an example, Fig. 2(a) depicts 30-year (1951-1980), monthly-mean dust events from surface observations at Dunhuang, China, located between the Taklimakan and great Gobi deserts. It is clear that the peak season of dust events over Dunhuang is during the boreal spring (March-May) and the lowest during the autumn (SeptemberNovember). Analyzing modern satellite observations, such as the TOMS Aerosol Index (e.g. Hsu et al., 1996), of a 2° x 2° region over Dunhuang during 1979-2000, the statistical features vary slightly, as illustrated in Fig. 2(b). The major discrepancy between these two temporal-spatial datasets is the seasonal trend of minimal dust events. Although satellite observations provide sizeable spatial coverage compared to that of surface in situ, they often experience an insensi-tivity of phenomena in the planetary boundary layer. Since the peak season of dust events over Dunhuang is evidently prolonged from the analyses of two completely different datasets, it is likely that during the winter season ground-based observations are more sensitive to aerosol loading in the atmosphere near the surface than those of satellite measurements.

Wang et al. (2004) presented a comprehensive overview of modern dust storm outbreaks in China and concluded that the most active geographic regions are (1) the Takli-makan desert, (2) the Hexi Corridor to the west Inner Mongolia Plateau, and (3) the central Inner Mongolia Plateau. Although the compiled results reveal seasonal changes of dust storm occurrence in different regions, the peak season is indicated during the boreal spring for all. In fact, the temporal distribution averaged over all regions in China closely mimics that of Dun-huang [see Fig. 2(a)]. However, with regard to the long-term trend of dust storm outbreaks in China, it is relatively divisive among studies reviewed by Wang et al. (2004). Overall, the highest frequency of dust storms occurred in the 1960s and 1970s; but they concluded that there were no significant statistical correlations between the frequency of dust storms and wind energy, or annual precipitation and evaporation. Besides the natural variability of weather/climate systems and other controlling factors, Wang et al. (2004) considered human activities, such as breaking up naturally wind-resistant surfaces and wiping out protective

Figure 2. (a) Surface observations of dust events at Dunhuang (40° 2' N, 94° 7' E), China, during 1951-1980, based on the criteria of horizontal surface visibility; (b) satellite analyses of dust events in the vicinity of Dunhuang (39-41°N, 93-95°E) during 1979-2000, anchored in deviations of spectral reflectance (i.e. Aerosol Index; Hsu et al., 1996) for indicating aerosol abundance in the column atmosphere; and (c) the annual trend (at 95% significance level) of dust storm distributions (Wang et al., 2004) in China during 1954-2000, utilizing records of 681 key stations operated by the China Meteorological Administration.

Figure 2. (a) Surface observations of dust events at Dunhuang (40° 2' N, 94° 7' E), China, during 1951-1980, based on the criteria of horizontal surface visibility; (b) satellite analyses of dust events in the vicinity of Dunhuang (39-41°N, 93-95°E) during 1979-2000, anchored in deviations of spectral reflectance (i.e. Aerosol Index; Hsu et al., 1996) for indicating aerosol abundance in the column atmosphere; and (c) the annual trend (at 95% significance level) of dust storm distributions (Wang et al., 2004) in China during 1954-2000, utilizing records of 681 key stations operated by the China Meteorological Administration.

vegetation from lands, to have played the most imperative role in the long-term trend of dust storm outbreaks in China over the last 50 years. As illustrated in Fig. 2(c), the four regions, which had experienced increases in the frequency of dust storm outbreaks, were largely caused by regional land desertification due to human activities.

1.2. Source/sink and transport pathway

Airborne dust particles can be found everywhere; the dust that falls in our backyard (the sink) may have originated in arid regions (the source) somewhere on Earth but traveled thousands of miles (the transport). Based on the analyses of 40 years of meteorological data (1960-1999 at 174 stations), Sun et al. (2001) concluded that springtime dust storms originating in China are highly associated with the activities of frontal systems and the Mongolian cyclonic depression. With some additions, their results on the routes/frequencies of cold air and dust storm outbreaks are summarized in Fig. 3.

From these 40-year statistics, the springtime dust storms in China were also revealed to have a peak occurrence in April, which was about three times those having taken place in either March or May. Overall, the frequencies of cold air outbreaks during springtime, originating from the west, north, and northwest, range from relatively comparable (Fig. 3, upper-left graph; Sun et al., 2001) to nearly doubled to quadrupled (respectively, 13.8%, 27.8%, and 58.6%; Gao et al., 2006), which all resulted in triggering dust storms from the great Gobi deserts along the Hexi Corridor and Inner Mongolia Plateau. However, only the western routes of cold air outbreaks frequently cause dust storms from the Taklimakan desert. Dust particles lifted from the great Gobi deserts (sum of frequencies > 80%; Fig. 3, upper-right graph) are commonly elevated up to 3 km, which is favorable for

Dust Storm China Graph

Figure 3. Statistical analyses of the source/sink regions and transport pathways for dust storm outbreaks in China (after Sun et al., 2001), overlying an EOS/MODIS product of a clear-sky, true-color composite image. Examples of dust aerosols near the source (lower-left picture, showing a passage of a dust front in Xinjiang province, western China) and over the sink (upper-right picture depicting dust fallouts over the Forbidden City in Beijing, China) regions are presented, as well as the routes/frequencies of cold air outbreaks (blue arrows/bars) and patterns/frequencies of transport pathways (brown arrows/bars).

Figure 3. Statistical analyses of the source/sink regions and transport pathways for dust storm outbreaks in China (after Sun et al., 2001), overlying an EOS/MODIS product of a clear-sky, true-color composite image. Examples of dust aerosols near the source (lower-left picture, showing a passage of a dust front in Xinjiang province, western China) and over the sink (upper-right picture depicting dust fallouts over the Forbidden City in Beijing, China) regions are presented, as well as the routes/frequencies of cold air outbreaks (blue arrows/bars) and patterns/frequencies of transport pathways (brown arrows/bars).

regional transport. Consequently, their impacts are limited to the regional scale, from the proximal Loess plateau and metropolitan Beijing to downwind areas of the Korean peninsula, Japan, Taiwan, and the nearby Pacific Ocean; whereas those dust aerosols originating mainly from the Taklimakan desert and rarely from the great Gobi deserts (total <20%) are frequently entrained to an elevation higher than 5 km and can be transported over long distances (e.g. ^5,000 km) by the prevailing westerly jet streams.

Atmospheric circulations over the Taklimakan desert are extremely complex due to the influence of bounding terrain [see Fig. 1(c)] which induces topographic channeling of the winds. The dominant moving direction of Taklimakan sand dunes, running up against the Tibetan plateau, is observed as either easterly or northeasterly, which clearly indicates the existence of prevailing low-level, easterly airflows. Thus, without a strong vertical lifting, the dust-laden atmosphere is poorly ventilated and the dust aerosols remain trapped in the enclosed basin. Recent lidar profiling of aerosol vertical distributions, such as the Geoscience Laser Altimeter System (GLAS) aboard the Ice, Cloud, and land Elevation Satellite (ICESat), is very useful for interpreting such cases. Figure 4(a) shows a Terra/MODIS red-green-blue composite image, depicting an outbreak of a Taklimakan dust storm. This dust storm began to be visible from space a few days before this image was taken. Approximately 6 hours prior to the closest satellite overpass, ICESat/GLAS captured this dust episode [Fig. 4(b)] when crossing the Taklimakan desert, indicated as a black line on the Terra/MODIS image (Fig. 4a). The dust layer was located about 3 km above ground, forming an arch across the desert but not lifted high enough to escape out of the basin. In addition, the dust layer in the southern

Figure 4. (a) Terra/MODIS red-green-blue composite image acquired on 14 March 2006, and (b) ICESat/GLAS vertical profile at 532 nm wavelength of a dust storm in the Taklimakan desert, with its corresponding geolocation marked as a black line on (a). Symbols on the transect line denote various mountain peaks in the Altai-Shan (circle and diamond), Tien-Shan (square), and Kunlun-Shan (star) ranges.

Figure 4. (a) Terra/MODIS red-green-blue composite image acquired on 14 March 2006, and (b) ICESat/GLAS vertical profile at 532 nm wavelength of a dust storm in the Taklimakan desert, with its corresponding geolocation marked as a black line on (a). Symbols on the transect line denote various mountain peaks in the Altai-Shan (circle and diamond), Tien-Shan (square), and Kunlun-Shan (star) ranges.

Earth Probe Toms Aerosol Index

Figure 5. Composite of TOMS Aerosol Index, a surrogate for aerosol loading, on 4-16 April 2001, depicting the long-range transport of Asian dust particles with a pathway across the Pacific Ocean and North America to the Atlantic Ocean. The arrows/dates indicate the size and location of the dust aerosols observed, with yellow-green-red colors in increasing order of density for dust loadings.

Figure 5. Composite of TOMS Aerosol Index, a surrogate for aerosol loading, on 4-16 April 2001, depicting the long-range transport of Asian dust particles with a pathway across the Pacific Ocean and North America to the Atlantic Ocean. The arrows/dates indicate the size and location of the dust aerosols observed, with yellow-green-red colors in increasing order of density for dust loadings.

portion of the arch against the Tibetan plateau was thick enough to prevent complete penetration of the laser beam, impairing estimations of the column dust loading and ground level detection. This also portrays the difficulties involved in monitoring, assessing, and analyzing complicated cases such as the Asian dust storm outbreaks when utilizing satellite sensors alone.

Nevertheless, satellite measurements provide dramatic results for the discovery that airborne dust particles can travel remarkably vast distances, as illustrated in Fig. 5, and induce various environmental impacts along their transport pathway. In the spring, Asian dust particles together with other anthropogenic and natural aerosols, once generated over the source regions, can be moved out of the boundary layer into the free troposphere. Sometimes riding with the westerly jet streams, they can travel thousands of miles across the Pacific into the United States, Canada, and beyond. As the infamous "2001 perfect dust storm" demonstrated, it took just less than a week to mobilize these dust plumes across the Pacific Ocean.

Part of these dust aerosols have been observed to subside in the Arctic, with some aloft in layers up to 10 km over Texas, and others linger over the NASA/Goddard Space Flight Center in Maryland (see NASA press release in April 2001).

2. Consequence

The processes of generating, transporting, and dissipating airborne dust particles are global phenomena — African dust regularly reaches the Alps; Asian dust seasonally crosses the Pacific into North America, and ultimately the Atlantic into Europe. One of the vital biogeo-chemical roles which dust storms play in Earth's ecosystem is the routine mobilization of mineral dust, as a source of iron (e.g. Meskhidze et al., 2005), from deserts into oceans for fertilizing the growth of phytoplankton — the basis of the oceanic food chain. Similarly, these dust-laden airs supply crucial nutrients for the soil of tropical rain forests, the so-called womb of life, which hosts 50-90% of the species on Earth. Historically, dust storm outbreaks are mainly natural events, but recent decades of increases in surface disturbances associated with anthropogenic activities worldwide may have altered considerably the net amount of airborne dust distributions. Unrestrained land use (e.g. deforestation, overgrazing, urbanization) coupled with long periods of drought can lead to massive deterioration of land cover, which can in turn increase the frequency of dust storm outbreaks.

More than 60% of the world's population resides in Asia, and China alone contains more than 20% (—1.3 billion) of the world's population. China relies primarily on agricultural products harvested from about 8-9% of Earth's arable and permanent cropland. In comparison, the United States, with a similar size of cropland, supports only a population of —0.3 billion. In addition, as one of the major sources of protein, China accounts for 33% of the global fish and seafood consumption (e.g. Pauly et al.,

2003), while domestic fisheries from the lakes, rivers, and coastal seas/oceans account only for 15% of the world's fish catch. As a result of meeting this demand, the production of aquaculture freshwater fish has leapt sharply, at the cost of reducing China's cropland. The balance of China's ecosystem is particularly delicate; thus, a fatal perturbation (e.g. drastic change of land cover and land use, depleted fisheries) can upset the balance and have an enormous impact on the global society due to the extremely large population size.

Synthesizing available data and records, Liu and Diamond (2005) presented and discussed in detail China's changing environment and socioeconomic challenges in the context of global interdependence. On the issues of natural disasters, it was noted that from AD 300 to 1949 northwestern China was exposed to major dust storms once every 31 years on average. Additionally, based on the statistics studied by the Chinese Academy of Sciences, the average frequency increased to about once per year from the 1950s to the 1990s. Since 2000, the average number of dust storms for the same region has escalated to 5-6 per year, including 8 fallouts in Beijing in 2006. In the drier northwest, overgrazing and overplowing severely degraded the vast natural grasslands in China, especially in Qinghai province and the Inner Mongolia Autonomous Region, two of the major source regions of dust storms. Liu and Diamond (2005) further concluded that 90% of China's grasslands have been degraded and declining at a rate of —1.5 x 104 km2 per year since the early 1980s. Moreover, the current status of desertification in China is very serious. The total area of desertification covers about 2.6 x 106 km2 (—27% of China's territory), of which 1.6 x 106 km2 of desertified land is caused by wind erosion. Studley (1999) estimated that the spreading rate of desertification is —2,460 km2per year. Similar figures were acknowledged by the Chinese authorities (i.e. Xinhua News), and appeared in a press conference held on 26 May 2001.

The consequence of desertification and topsoil erosion in China, in terms of dust storm outbreaks, has had a worldwide impact and appeared frequently in the springtime headline news of various media. A particular example is the catastrophic losses of human lives and property, crops, and livestock in the 5 May 1993 dust storm (Yang, 2001). The ways that airborne dust particles affect humankind's sustainable developments are wide-ranging and can be classified broadly under societal impact and scientific significance.

2.1. Societal impact

With massive amounts of dust lifted from China's desert regions and injected into the atmosphere, these dust storms often affect daily activities in dramatic ways: pushing grit through windows and doors, forcing people to stay indoors, causing breathing problems, reducing visibility and delaying flights, and by and large creating chaos. Essentially, their long-lasting consequences are:

• Busting crop yields. Airborne dust particles not only decrease the crop yields through the reduction of solar radiation reaching the plants for photosynthesis (Chameides et al., 1999), but also contribute to the spread of agricultural diseases, such as bacteria and fungi of plant pathogens that are primarily devastating to rice and wheat — the staple food grains in Asia. Given the projections of limited and declining cultivable croplands and a steadily rising population with increasing per capita food consumption, this absolutely threatens food security in China in the coming decades (Brown, 2002).

• Posing health risks. Toxic effects on people near the source/sink and along the pathway of dust storms are evident, since dust contains aluminum, zinc, iron, and other trace metals, which irritate the eyes and respiratory system. Two to three days after dust storm events, the escalation in hospital admissions for asthma (Yang et al., 2005a) and cardiovascular disease (Yang et al., 2005b) is prominent and statistically significant. Furthermore, these adverse effects of air pollution including inhalable dust aerosols on human health are blamed for increases in bronchitis, meningitis, and even premature mortality (e.g., Xu et al., 1998) in the Asia-Pacific region. According to the World Bank's 2000 Annual Review, the health costs in major Asian cities can reach as much as 15-18% of urban income. • Causing economic hazards. During its life cycle — generated at the source region, transported along the pathway, and dissipated over the sink area — a dust storm with an average strength could have numerous direct economic impacts (e.g. injuring people and destroying properties, damaging crops and forests, suffocating livestock, reducing semiconductor yields), as well as influence the economy in many indirect ways (e.g. increasing health costs, depleting fisheries due to a triggered red tide, disrupting transportation and communication, shortening the functional life of hydroelectric dams and river channels). The combined economic loss was estimated by the United Nations Environment Program to be in the range of $6.5-9.1 billion each year — roughly equivalent to 0.6-0.9% of China's gross domestic product in 2001.

2.2. Scientific significance

Since the early 1970s there has been a series of successful launches of Landsat satellites capable of global observations, so it is no wonder that dust science at the planetary scale is on the rise. This momentum has been fueled by the dawn of the Earth Observing System (EOS) era in the late 1990s — currently, there are 16 (and counting), active polar-orbiting satellites that provide unprecedented views of the size, scale, distribution, and movement of the dust storms from space. The unique vantage provided by satellites in detecting and tracking the progress of airborne dust has helped to shed new light on how humankind has affected the development and magnitude of these storms. A synergy of multisatellite observations suggested an increase in both the intensity and the frequency of Asian dust storm outbreaks that was paralleled by the spatial and temporal scales of manmade development, and subsequent change in the qualities of the land surface, occurring in northwestern China.

From a macro-scale perspective, satellite measurements of dust characteristics offer vital information for determining how dust particles affect the weather and climate by redistributing solar energy within Earth's atmosphere and by changing the thermal contrast between land and ocean. When interacting with sunlight, airborne dust particles not only absorb solar radiation but also reflect it back to space. Generally, this results in a net warming in the dust-laden atmosphere and a net cooling on Earth's surface. However, the magnitude of its cooling over the ocean surface differs from that over the land. Exactly how dust particles modify Earth's radiation budget depends on their color, size, shape, and chemical composition. Furthermore, airborne dust particles may have very complex interactions with atmospheric waters and aerosols from a micro-scale perspective. In some instances, atmospheric aerosols including dust particles serve as nuclei for condensing raindrops, eventually leading to enhancement/acceleration of precipitation processes (e.g. Shepherd and Burian, 2003), but in other cases those aerosols just stifle precipitation (e.g. Rosenfeld, 2000). It is hardly surprising to find that the interactions of dust particles, among other types of aerosols, with the key factors of meteorology (water vapor supply), dynamics (diffusion, collision, and coalescence), and microphysics (water and ice nucleation), can be very complicated (e.g. Li and Yuan, 2006).

To understand the profound, complex, and far-reaching dust impacts on the Earth-atmosphere system, it is of paramount importance to assimilate the spatial and temporal variability of airborne dust properties. Given the daunting diversity of airborne dust properties, a combined observational and theoretical approach is required to better understand and quantify the effects of dust. In doing so, many questions arise:

• Could these increasing airborne dust particles, together with anthropogenic pollutants, drastically alter the regional cloud distributions and hydrological cycles through aerosol direct and indirect effects (variability and forcing)?

• Would essential fisheries in this region be critically impacted by dust storms through the influence of the nutrient deposition pattern and extent, in terms of primary productivity of plankton (response and consequence)?

• How could we strengthen our ability, through a better understanding of dust properties and interactions with regional meteorology, to issue early warnings of dust storms and of adverse health effects on humans (prediction)?

• To what extent could we assess the effectiveness of increasing vegetation/trees (e.g. a reforestation project on the outskirts of Beijing) in preventing further desertification (feedback and prediction)?

3. Dust Properties

Solar radiation is the sole large-scale source of diabatic heating that drives the weather and climate system on planet Earth. The emission of terrestrial radiation back to space keeps the planet in balance to make it habitable for all forms of life. Aerosols play an important role in modifying the distributions of solar and terrestrial radiation (IPCC, 2001). Four major types of aerosols — dust particles, biomass-burning smoke, air pollutants, and sea salts — commonly occur in the atmosphere. How light is scattered, absorbed, and emitted by aerosols depends critically on their physical and chemical properties, including refractive index, species, mixture, hygroscopicity, size distribution, shape, and orientation. A thorough understanding of aerosol properties, as well as their temporal and spatial distribution, is imperative for comprehending how Earth's atmosphere maintains its current state of equilibrium and how anthropogenic activities can potentially ruin that balance. Information obtained from coordinated ground-based (temporal scale) and spaceborne (spatial scale) measurements will allow scientists to study in detail the properties of dust particles from sources to sinks and along transport pathways.

3.1. Ground-based observation and analysis

Spaceborne remote sensing observations are often plagued by contamination of surface signatures. Thus, ground-based in-situ and remote sensing measurements, where signals come directly from the atmospheric constituents, the Sun, and/or the Earth-atmosphere interactions, provide additional information for comparisons that confirm quantitatively the usefulness of the integrated surface, aircraft, and satellite datasets. Under the auspices of the International Global Atmospheric Chemistry Program, a most comprehensive field campaign in East Asia, the Aerosol Characterization Experiment-Asia (ACE-Asia; Huebert et al., 2003) was conducted in the spring of 2001 and beyond to study the Asian dust and pollutant aerosols. During the ACE-Asia intensive observation period (IOP), many surface sites, together with three aircraft, two research ships, and numerous EOS satellite overpasses, made simultaneous measurements of aerosol chemical, physical, optical, and radiative properties under a variety of environmental conditions. Subsequent to the ACE-Asia IOP, a few surface sites (e.g. the Gosan site in S. Korea, the Mt. Bamboo site in Taiwan) continued to operate so as to acquire in situ and column-integrated aerosol properties to assess their spatial and temporal variability.

Near the source regions of dust storm outbreaks, as depicted in Fig. 6(a), a subset of the NASA SMART(Surface-sensing Measurements for Atmospheric Radiative Transfer)-COMMIT (Chemical, Optical and Microphysical Measurements of In situ Troposphere) facility was deployed at the Dunhuang site, located

Figure 6. (a) An early-stage SMART-COMMIT instrumental setup at Dunhuang, China, during the ACE-Asia campaign on a dusty day; (b) a newly established LABS instrumental setup on the building rooftop; and (c) distant view of the LABS facility located at an altitude of km near the top of Lulin mountain, Taiwan.

Figure 6. (a) An early-stage SMART-COMMIT instrumental setup at Dunhuang, China, during the ACE-Asia campaign on a dusty day; (b) a newly established LABS instrumental setup on the building rooftop; and (c) distant view of the LABS facility located at an altitude of km near the top of Lulin mountain, Taiwan.

19:40 UTC

19:40 UTC

Figure 7. On 28 April 2001, a fierce dust storm passed right over the Dunhuang site. A total-sky imager captured a series of sky conditions (upper panel) for a nearly clear sky (for comparison), a few hours before the storm (16:32 UTC), and the other four images quick succession as the storm passed over. The vertical distribution of dust properties (e.g. aerosol extinction profile) is measured by a micropulse lidar (lower-left panel) during transport. Two photos contrast the changes of airborne dust particles, before (lower-center, 1:30 p.m. local time) and 1 hour after the storm passed (lower-right).

Figure 7. On 28 April 2001, a fierce dust storm passed right over the Dunhuang site. A total-sky imager captured a series of sky conditions (upper panel) for a nearly clear sky (for comparison), a few hours before the storm (16:32 UTC), and the other four images quick succession as the storm passed over. The vertical distribution of dust properties (e.g. aerosol extinction profile) is measured by a micropulse lidar (lower-left panel) during transport. Two photos contrast the changes of airborne dust particles, before (lower-center, 1:30 p.m. local time) and 1 hour after the storm passed (lower-right).

between the Taklimakan and Gobi deserts. SMART-COMMIT (see http://smart-commit. gsfc.nasa.gov for more details) is a mobile laboratory, consisting of many commercially available and in-house-developed remote sensing instruments, as well as a variety of in situ probes, for measuring aerosol and precursor properties and meteorological parameters. Likewise, a suite of similar instrumentation, with additional sensors particularly for cloud water chemistry, was deployed for ACE-Asia and ADSE (Asian Dust Storm Experiment, 20012004) at the sink areas of dust aerosols in northern Taiwan, known as the Mt. Bamboo site, at an ~1.1 km altitude. Furthermore, recognizing the vital environmental impacts of aerosol longrange transport, the Lulin Atmospheric Background Station (LABS) was established and has been operational since 13 April 2006, as shown in Figs. 6(b) and 6(c). Situated near the mountaintop (~3km height) in central Taiwan and equipped with advanced/refined instrumentation from the Mt. Bamboo site, measurements from the LABS (see http://labs.org.tw for more details) are observing key parameters representing the free troposphere after the ACE-Asia and ADSE projects.

Among many research topics, Fig. 7 illustrates, for the first time, how dramatically a fierce storm generated a dense blanket of dust, and was observed by a suite of sophisticated ground-based instruments and spaceborne sensors (not shown). On 28 April 2001, a total-sky imager captured the entire passage of this fast-approaching dust storm at ~3p.m. local time right over the ACE-Asia Dunhuang site. Daily vertical distributions of dust aerosols (Fig. 7, lower-left panel) were documented by a micropulse lidar, form which an aerosol extinction profile can be retrieved quantitatively. Dense dust particles completely obstructed the lidar backscattering signals immediately after the dust front passage (~5 minutes), followed by the snowfall. Also shown in Fig. 7 is a photo (lower-center) taken at 1:30 p.m. local time for the approaching dust storm, while another photo (lower-right) demonstrates the dense blanket of dust in the air approximately 1 hour after the storm passed. The yellowish color clearly indicates different chemical compositions (e.g. iron content) in the dust particles, as compared to those of brownish Saharan dust. From satellite observations, this dust storm began to form on 27 April 2001, intensified on 28 April, moved eastward on 29-30 April, and dissipated on 1 May.

During transport, airborne dust particles can interact with anthropogenic sulfate and soot aerosols from heavily polluted urban areas. Added to the complex effects of clouds and natural marine aerosols, dust particles reaching the downwind and sink areas can have drastically different properties than those from the sources. In ACE-Asia, micrographs of airborne dust particles clearly reveal changes of their compositions during transport. Dust particles sampled at the Dunhuang site (source regions) contain predominately silicates, with additional clay minerals, carbonates, feldspars, and gypsum present, while they become dirtier by the time passing over polluted areas. Aircraft observations of dust micrographs by J. Anderson (Arizona State University) show that many different forms of soot balls and nonsoot carbonaceous particles aggregate with the mineral dust particles.

When these dust particles are transported downwind to the sink areas associated with frontal activities, the dry/wet removal processes involving interactions with clouds, biomass-burning aerosols, and local pollutants further complicate atmospheric composition and tro-pospheric chemistry. During ADSE (e.g. Lin and Peng, 1999), cloud waters were collected hourly from the Mt. Bamboo site at an altitude of ~1.1 km, which was frequently immersed in liquid water clouds due to either frontal passage or topographic lifting. Considering the location and altitude of the Mt. Bamboo site and the prevailing northeast-monsoon winds, cloud waters sampled upwind should not be contaminated by megacity pollutants emitted from Taipei. Subsequently, the collected cloud waters were measured to obtain their acidity (in terms of pH values), and conductivity, and were analyzed for ion concentrations of Cl-, NO- , SO^-, NH+, Na+, K+, Mg2+, and Ca2+ using ion chromatography (e.g. Lin et al., 1999) to determine cloud chemical composition.

Figure 8 shows an assessment of dust and other anthropogenic aerosols influencing cloud chemistry by means of colored cloud waters (unusually brown-to-black), compared to a normally transparent color from natural clouds. Furthermore, the pH value of one sample

Bluetooth Controller

Figure 8. On 22-23 March 2002, surface measurements indicated variations of cloud water chemistry due to a dust storm event over the Mt. Bamboo site, Taiwan. (a) Cloud water collector; (b) close-up view of contaminated cloud water strings; (c) hourly collected cloud water; (d) conductivity and pH measurements; and (e) hourly filter papers (date/local-time in white) revealing changes of color appearance and pH values (gray-to-black numbers).

Figure 8. On 22-23 March 2002, surface measurements indicated variations of cloud water chemistry due to a dust storm event over the Mt. Bamboo site, Taiwan. (a) Cloud water collector; (b) close-up view of contaminated cloud water strings; (c) hourly collected cloud water; (d) conductivity and pH measurements; and (e) hourly filter papers (date/local-time in white) revealing changes of color appearance and pH values (gray-to-black numbers).

reached as high as 7.89, alkalized typically by the calcium ion (Ca2+), a major composition of dust particles. By contrast, the pH values of cloud waters were generally 3.0-4.5 at the Mt. Bamboo site. The conductivity of this sample was also measured to as high as 962 ¡iQ,-1cm-1, ~10 times more than that of rainwater collected in Taipei, indicating exceedingly contaminated conditions. The time series filter papers of cloud water [Fig. 8(e)] clearly depict the varying degrees of aerosol contaminations: prevailing north-northeast winds carried dust-dominant aerosols (pH values 6), but abruptly changed (at 21-22 local time on 23 March) to south-southwest winds with biomass-burning and megacity aerosols (pH values < ~ 5) respectively from Southeast Asia and locally. Essentially, the brown and black colors represent respectively the dust and biomass-burning aerosols. The former tends to alkalize the cloud waters, causing lower acidity in comparison with the CO2-equilibrated pH value of 5.6. However, biomass-burning and megacity aerosols generally consist of black and organic carbons and associated acidic substances (e.g. SO=), resulting in increasing acidity of the cloud waters.

From satellite observations combined with a five-day backward trajectory analysis, this dust storm outbreak started in Inner Mongolia on 18 March 2002, moved eastward on 19-20 March, and blanketed the entire Korean peninsula (the highest measured PMio concentration exceeded 3,000 ^gm-3 and thus the instrumental detection limit) on 21-22 March. This dust outflow, associated with a frontal passage, reached northern Taiwan in the late evening of 22 March and lasted until midday of 24 March. Meanwhile, widespread biomass-burning activities in Southeast Asia had also been observed since 19 March. During the period of this event, ADSE micropulse lidar of NASA/MPLNET detected two layers of strong backscattering signals (i.e., lower-level airborne dust particles advected from the north-northeast and higher-level, ^2-3 km, biomass-burning aerosols from the south-southwest). However, the frontal system played an important role in the mixing of dust/biomass-burning aerosols with clouds. The complex interactions among aerosols, clouds, and atmospheric constituents are an important topic to be studied of atmospheric composition and tropospheric chemistry in ACE-Asia.

Radiative forcing is an area of keen scientific interest, because it is a key parameter in understanding the perturbations that drive the weather and climate system. To quantify the energetics of the surface-atmosphere system, accurate surface measurements of broadband shortwave (0.3-2.8 ¡m) and longwave (4.050 ¡m) irradiance by flux radiometers are required. By combing irradiance measurements from satellites at the top of the atmosphere and those from ground-based radiometers with aerosol optical thickness, the radiation budget of the surface-atmosphere system can be determined over an extensive area (e.g. Hansell et al., 2003). However, due to the temperature gradients between the filter dome and detector in flux radiometers, surface measurements must be corrected to account for the thermal dome effect (e.g. Ji and Tsay, 2000), which may range from 5 to 20 watts m-2 in magnitude, depending on the state and condition of the atmosphere. Applying the synergy of the surface/satellite multisensor approach, Hsu et al. (2000) demonstrated that the presence of Saharan dust results in a net cooling over ocean and a net warming over land. However, care should be taken in screening out cloud contamination when one extends this approach to studying the radiative forcing of Asian dust, which is generally associated with cloudy, moisture-laden weather fronts. On the other hand, with radiance sensors acquiring spectral or narrowband measurements in the visible, shortwave-infrared, longwave-infrared, and microwave regions and lidar backscattering intensity at the surface, accurate knowledge of the atmospheric aerosols and constituents can be extracted, such as aerosol optical thickness and corresponding vertical profile, columnar size distribution, column water vapor/liquid amount, and ozone abundance. These retrieved parameters can be used to initialize the atmospheric aerosol profile in the forward calculation of radiation models, and to evaluate the results of numerical modeling studies.

3.2. Satellite monitoring and retrieval

Among all atmospheric properties to be monitored and retrieved from space, tropospheric aerosols are especially important, since they have a relatively short lifetime with large temporal and spatial variations. Atmospheric aerosols affect various aspects of solar and terrestrial radiative transfer in spectral (A), spatial (x, y, z), angular (0,4*), and temporal (t) domains; in turn, the operational satellites use one or more of these four aspects for monitoring, assessing, and retrieving the aerosol properties and effects.

The longest record of aerosol observations from satellites can be dated back to late 1978 with TOMS, and continues to the present time of Aura/OMI (Ozone Monitoring Instrument), in the context of an aerosol index by inverting spectral measurements. The TOMS aerosol index (Hsu et al., 1996) is determined from a pair of ultraviolet spectra that respond negligibly to ozone absorption but strongly to Ray-leigh scattering (e.g. 340, 360, and 380 nm). Because the spectral reflectivity of cloud and surface varies, weakly in the ultraviolet wavelengths, the TOMS aerosol index can be used unambiguously to differentiate aerosols from clouds, and to detect absorbing aerosols over arid and semiarid surface. Unlike the thermal-contrast method using infrared spectra, the detection of mineral aerosols by TOMS is not susceptible to water vapor absorption and surface temperature variation. As illustrated in Fig. 5, the TOMS aerosol index can be used to monitor the evolution of dust storms after dust particles are lifted from the source regions. Also, the frequency statistics of the aerosol index can be utilized for obtaining the source information attributed to multiple types of aerosols, such as airborne dust particles, biomass-burning smoke, and air pollution. However, since the aerosol index represents a measure of how atmospheric molecules intervene with absorbing par-ticulates and the concentration of molecules strongly depends on atmospheric pressure (e.g. Penndorf, 1957), the resulting signals are quite sensitive to the height of particulates residing in the atmosphere. Thus, at the current stage the aerosol index is considered to be an extremely valuable qualitative product.

Although the TOMS aerosol index provides much information about absorbing aerosols, the once-per-day observation limits its timely applications, such as issuing near-real-time warnings of dust storm outbreaks. Utilizing the spectral and temporal aspects of a geostationary satellite, Liu and Lin (2004) successfully developed an automatic operational system for Asian dust storm detection and monitoring, based on measurements of GMS-5 S-VISSR (Stretched-Visible and Infrared Spin Scan Radiometer) from the Japan Meteorological Agency. The efficacy of GMS-5 S-VISSR observations for timely dust storm monitoring is made possible by the broad spectral channels (visible — 0.73 ¡m; water vapor — 6.75 ¡m; infrared split-window — 11-12 ¡m) and high temporal resolutions (hourly).

Figure 9(a) depicts the scatter plot for GMS-5 S-VISSR digital counts of visible versus infrared channels, which clearly identify distinct atmospheric and environmental aspects such as dust, cloud, land, and ocean surface. With wide spatial coverage and high temporal resolution of the geostationary satellite, the dust-free background data can be established before a dust storm occurs. Thus, the source regions of the dust storm and dust-affected areas can be determined through the variances between observed and background data. However, comparable

Aerial 255 Jpg

Figure 9. GMS-5 S-VISSR measurements at 02:40Z 7 April 2001: for (a) a scatter plot of digital counts between visible and infrared channels (note the higher the counts the brighter for visible or the colder for infrared scene) and (b) dust storm detection illustrated by the Sand Index gray scale, in which light gray stands for dust-free and white for cloudy areas, while an increasing order of the gray scale represents the abundance of the atmospheric dust loading.

Figure 9. GMS-5 S-VISSR measurements at 02:40Z 7 April 2001: for (a) a scatter plot of digital counts between visible and infrared channels (note the higher the counts the brighter for visible or the colder for infrared scene) and (b) dust storm detection illustrated by the Sand Index gray scale, in which light gray stands for dust-free and white for cloudy areas, while an increasing order of the gray scale represents the abundance of the atmospheric dust loading.

albedo values between dust particles and cloudy neighborhoods, especially around low cloud-cover regions, may lead to major misdetections. To overcome this difficulty, the split-window technique (brightness temperature difference of 11-12 ¡m) is employed to provide water vapor information, since the main difference between dusty and cloudy scenes is the water vapor content. An example for detecting and monitoring the formation of a dust storm and delineating dust-affected regions is presented in Fig. 9(b). These results clearly demonstrate the superiority by using both infrared channels of GMS-5 to mitigate misdetections of dust-cloud neighboring regions.

To assist policymakers in making accurate and timely decisions, this automatic detection/ monitoring system is readily embedded in the GMS-5 satellite receiving system for expediting the provision of information about dust storm outbreaks, evolution, and affected areas. There are still two caveats: the low dynamical range of the S-VISSR visible channel leaves little sensitivity for differentiating dust-contaminated clouds from clean clouds, and the usage of the visible channel prevents its nighttime applications.

Aerosol optical thickness (0 < AOT < ro), an optical measure of aerosols loading, is the most quantitative and fundamental property in describing the consequence of aerosols interacting with spectral light. The earliest reliable retrievals of AOT from space, as shown in Fig. 10(a), have been using two spectral radiance measurements of NOAA/AVHRR (Advanced Very High Resolution Radiometer) since 1983, but have only been available over global oceans (Geogdzhayev et al, 2002). Due to limited spectral channels, the other parameter retrieved from AVHRR is the Angstrom exponent (a), which indicates the size groups of dominant aerosols — the larger the a, the smaller the size (e.g. a = 4, for molecules; a & ± 0, for dust particles). Starting from the EOS era in the late 1990s, measurements from advanced satellite sensors with wide spatial coverage and multispectral channels, as well as innovative algorithms for analyses, have created a new horizon for retrieving aerosol properties. An excellent example is the MODIS sensors aboard

Figure 10. Examples of (a) the monthly mean of aerosol optical thickness at 0.55 ¡m wavelengths retrieved over global oceans using NOAA/AVHRR measurements (Geogdzhayev et al., 2002); (b) same as in (a), but retrieved over global vegetated land and open oceans (no retrievals over bright-reflecting surfaces, denoted in black over land) using EOS/MODIS measurements.

Figure 10. Examples of (a) the monthly mean of aerosol optical thickness at 0.55 ¡m wavelengths retrieved over global oceans using NOAA/AVHRR measurements (Geogdzhayev et al., 2002); (b) same as in (a), but retrieved over global vegetated land and open oceans (no retrievals over bright-reflecting surfaces, denoted in black over land) using EOS/MODIS measurements.

both the NASA/EOS Terra and Aqua satellites, which are successfully making near-global measurements daily permitting the retrieval of spectral AOT and aerosol size parameters over both land and ocean (e.g. Remer et al., 2005). Figure 10(b) depicts one of the current MODIS AOT products over global vegetated land and open oceans.

New information obtained from the advanced sensors now allows scientists to better understand the optical and microphysical properties of mineral dust which help to improve the mapping and prediction of dust storm outbreaks. In particular, measurements made by different satellite sensors passing overhead at different times have been instrumental in studying the creation and evolution of dust plumes over time. However, as illustrated in Fig. 10(b), aerosol properties in the vicinity of major desert regions, where dust storms frequently originate, are still largely missing due to large uncertainties in surface emissivity and/or reflectivity, as well as uncertainties in vertical profiles of aerosol and water vapor that are required in the conventional retrieval algorithms. Although Terra/MISR (Multi-angle Imaging SpectroRa-diometer) utilizes the spectral and angular aspects to overcome this difficulty of retrieving aerosol properties over bright-reflecting surface (Diner et al., 2005), its relatively narrow spatial coverage significantly impacts the resulting statistics of airborne dust properties near the source regions (Hsu et al., 2006).

Aerosol retrievals over bright-reflecting surfaces (e.g. airborne dust particles near desert regions) have been a challenging problem ever since the applications of satellite remote sensing. Essentially, the radiance received by a satellite sensor (or apparent radiance) comprises the direct scattered/emitted radiance by the scene and the path radiance that represents the contribution of scattered/emitted radiance directly and/or diffusely by the atmosphere and surface. Over the solar spectral wavelengths, the presence of aerosols in the atmosphere would practically brighten the scene over dark surfaces and darken the scene over bright-reflecting surfaces. The principle of the newly developed Deep Blue algorithm (Hsu et al., 2004) for retrieving aerosols over bright-reflecting surfaces takes advantage of the darker properties of such a surface at the blue spectra, as depicted in Fig. 11. For the first time, the Deep Blue retrievals, with optimal spectral wavelengths from the past, current, and future MODIS-like sensors (e.g. GLI, SeaWiFS, MODIS, VIIRS), can provide comprehensive aerosol properties that permit scientists to quantitatively track the evolution of dust and fine-mode anthropogenic aerosols from source to sink regions, with only

Figure 11. (a) Nadir spectral reflectance acquired over the Sahara desert (top of atmosphere — thin line; without molecular scattering — thick line) and over the surface of Gobi deserts (thick, dashed line), and (b) aerosol efficiency index, indicating the percentage change of the apparent spectral reflectance in comparing a Rayleigh (aerosol-free) atmosphere to a Rayleigh atmosphere containing aerosols (the higher the values, the more sensitive the spectra in detecting the presence of aerosols). Two dust models (D2 for regular and D1 for extremely absorbing dust aerosols) and two operational satellite-viewing geometries (0o for nominal and 60o for limb) are used to cope with the sensitivity ranges of dust effects. The optimal and currently available Deep Blue channels for aerosol retrievals are indicated as vertical bars (e.g. 412, 470, and 670 nm).

Figure 11. (a) Nadir spectral reflectance acquired over the Sahara desert (top of atmosphere — thin line; without molecular scattering — thick line) and over the surface of Gobi deserts (thick, dashed line), and (b) aerosol efficiency index, indicating the percentage change of the apparent spectral reflectance in comparing a Rayleigh (aerosol-free) atmosphere to a Rayleigh atmosphere containing aerosols (the higher the values, the more sensitive the spectra in detecting the presence of aerosols). Two dust models (D2 for regular and D1 for extremely absorbing dust aerosols) and two operational satellite-viewing geometries (0o for nominal and 60o for limb) are used to cope with the sensitivity ranges of dust effects. The optimal and currently available Deep Blue channels for aerosol retrievals are indicated as vertical bars (e.g. 412, 470, and 670 nm).

moderate sensitivity to uncertainties in aerosol plume height (Hsu et al., 2006).

An example of MODIS Deep Blue AOT retrievals on 6 April 2001 is given in Fig. 12(c), which clearly reveals extensive dust plumes occurring over Mongolia, the Taklimaken desert, and Inner Mongolia. The corresponding Terra/MODIS red-green-blue image (10:30 a.m. overpass) and current MODIS-operational AOT products are also shown in Figs. 12(a) and 12(b), respectively. Because of the bright su

Was this article helpful?

0 0

Post a comment