Journal of Geophysical Research, Vol.102, NO. D14, 16889-16909, 1997

Characterization of troposphereic aerosols over the oceans with the AVHRR advanced high resolution radiometer optical thickness product

Rudolf B. Husar
Center for Air Pollution Impact and Trend Analysis
Washington University
St. Louis, MO 63130-4899

Joseph M. Prospero
University of Miami, RSMAS
Miami, FL 33149

Larry L. Stowe
World Weather Building
Washington, DC 20233



1.0 Introduction

2.0 The NOAA AVHRR Aerosol Product

3.0 The AVHRR Aerosol Record: July 1989 - June 1991

3.1 General Features of Aerosol Distribution Patterns

3.2 Atlantic Ocean

3.3 Indian Ocean, Arabian Sea and the Bay of Bengal

3.4 Pacific Ocean

4.0 Conclusions

5.0 References

Abstract. The National Oceanic and Atmospheric Administration (NOAA) Advanced Very High-Resolution Radiometer (AVHRR) is a polar orbiting satellite that provides information on aerosol distributions based on backscatter radiation measurements that yield a measure of the "radiatively equivalent" aerosol optical thickness (EAOT) over the oceans. Seasonally-composited EAOT data for the period July 1989 to June 1991 reveal many spatially-coherent plume-like patterns that can usually be interpreted in terms of known (or reasonably-hypothesized) sources in association with climatological wind fields. The largest and most persistent areas of high EAOT values are associated with wind-blown dust and biomass burning sources; especially prominent are sources in Africa, the middle East and the Asian sub-continent. Prominent plumes over the mid-latitude North Atlantic are attributed to pollution emissions from North America and Europe. Large plumes attributed to pollution aerosols and dust from sources in Asia are clearly visible over the western and central North Pacific. On a global scale, the annually averaged northern hemisphere EAOT values are about 1.7 times greater than those in the southern hemisphere. Considering each hemisphere separately, EAOT values in the summer are about twice those in the winter; within the mid latitude band 30o-60o (i.e., where anthropogenic emissions are greatest), the summer/winter ratio is about three. The temporal variability of monthly mean EAOT in specific ocean regions often shows characteristic seasonal patterns that are usually consistent with aerosol measurements made in the marine boundary layer. Nonetheless, there are many features in the EAOT distributions that can not be readily interpreted at this time. The AVHRR EAOT distributions demonstrate that satellite products can serve as a useful tool for the planning and implementation of focused aerosol research programs and that they will be especially important in studies of climate-related processes.

1.0 Introduction

Tropospheric aerosols play an important biogeochemical role in many earth processes; they can also affect human health, visibility and climate. Aerosols can impact climate by radiative processes, directly through the scatter and absorption of solar radiation and indirectly through cloud effects. Recent modeling of direct effects indicates that anthropogenic sulfate could scatter a substantial amount or radiation back to space [Charlson et al., 1992] thereby having a cooling effect on climate which appears to be substantiated by observed temperature trends [Karl et al., 1995]. Until recently, research has primarily focused on anthropogenic sulfate aerosols; however, other types of aerosols could be important, among them soil dust and the products of biomass burning [Penner et al., 1994; Penner, 1995; Andreae, 1995; Duce, 1995; Prospero, 1996a,b]. At present the estimates of direct aerosol forcing are highly uncertain because of many unknown factors about aerosol properties [Kiehl and Briegleb, 1993; Penner et al., 1994; Andreae, 1995; Karl et al., 1995; Lacis et al., 1995; Taylor and Penner, 1994] and their temporal and spatial variability over the earth, especially over the oceans where so little data is available. The concentration, composition and physical properties of particles in the marine atmosphere can vary greatly, depending on the distribution of sources, the controlling meteorological processes in the source regions, the large-scale wind systems involved in transport and, finally, the removal processes that act on the particles and deposit them in the ocean or carry them to other land masses. Consequently it is difficult to characterize aerosols solely on the basis of sporadic in situ measurements.

Satellites can be used to map aerosol distributions because aerosol particles, when illuminated by the sun, scatter a fraction of the solar radiation back to space. At the simplest level, aerosol transport events can be detected by visible inspection of imagery [e.g., Ferrare et al., 1990]; aerosol patches appear as diffuse gray areas which are clearly different from cloud which has a brighter and more textured appearance. The quantitative interpretation of satellite-sensed aerosol backscatter in terms of aerosol optical thickness (AOT) requires an accurate knowledge of a number of factors. Especially important are the aerosol angular scattering (phase) function, which is strongly dependent on the aerosol size distribution and chemical composition, and the albedo of the underlying surface. Retrievals are especially difficult over land surfaces which have complex and variable radiative properties. The process is simplified over the oceans because the ocean surface has a relatively low and constant albedo [Griggs, 1983; Durkee et. al., 1986, 1991]. Satellite-sensed AOT measurements over the ocean are generally in reasonable agreement with concurrent surface-based measurements of AOT as measured with upward-looking radiation instruments such as sun-photometers [Durkee et al., 1991; Villevalde et al.,1994; Ignatov, et al., 1995]. In this paper we present AOT values over the ocean.

2.0 The NOAA AVHRR Aerosol Product

The NOAA National Environmental Satellite Data and Information Service (NESDIS) has an operational program that routinely derives and archives estimates of AOT for the global ocean using polar-orbiting meteorological satellites. A detailed description of the history of the AVHRR program, the evolution of the aerosol retrieval algorithm and the analysis procedure is found in Stowe et al. (this volume). In this paper we present only a brief overview of the procedures used for the data presented herein.

The AOT is estimated from backscatter radiation measurements made at an effective wavelength of 0.63 m. The NOAA/11 AVHRR aerosol retrieval algorithm is based on a simple Junge aerosol size distribution (size parameter, =3.5) and a real index of refraction of 1.5. The algorithm corrects for Rayleigh scattering by atmospheric gases, for ozone absorption using global mean climatological ozone distributions, and for surface reflectance by assuming a constant global Lambertian ocean reflectance of 1.5%. The AOT distributions are presented as 1ox1o composites of the retreived data in 2x2 arrays of 4-km global area coverage pixels over the oceans. AVHRR data are available through the National Climatic Data Center (Asheville, NC).

To avoid solar specular reflection from the ocean surface, measurements are made only on the side of the orbit away from the sun, excluding data within a 40o half angle interval on the specular ray. To minimize the possibility of errors over coastal waters where water color and reflectance may be variable (because of sediment load from rivers, marine organisms, pigments, etc.) data from coastal waters are rejected. In heavy seas, the bright foam from whitecaps could cause spurious backscatter that could be interpreted as aerosol. During the data processing, this possible source of interference was not considered. However the AVHRR product (Fig. 1) does not show substantially increased EAOT values in the winter hemisphere where winds and whitecap coverage [Erickson et al., 1986] are expected to be highest; this suggests that whitecaps do not present a major problem.

Cloud screening procedures are based on the fact that the cloud reflectance is high and relatively constant across much of the visible and infra-red spectrum; in contrast, aerosol backscattering is much stronger in the visible than in the near and far infrared [McClain, 1989]. Pixels containing clouds are identified and removed by comparing the observed signal in the near-IR with that expected in the absence of clouds. The resulting estimated aerosol pattern may be spatially and temporally biased according to the fraction of the data that is retrieved for a specific region, a factor that is largely determined by cloud-cover statistics, by the equator crossing time of the satellite, and by not scanning the solar side of each orbit.

Validation tests [Ignatov et al., 1995] show good agreement between the AVHRR EAOT measurements and ground based AOT measurements. The good agreement is somewhat surprising, given the relatively simple algorithm that is used for the retrievals and the complex mixture of pollutant and natural aerosols that is found over many ocean regions. Pollutant aerosols and biomass burning products lie mostly in the submicron diameter size range. Mie scattering calculations (R. Husar, unpublished data) show that, for aerosols having a narrow size distribution at the low end of the submicron size range, the phase function is nearly symmetrical with strong forward and backward lobes. In contrast, for mineral dust and sea-salt, aerosols which have mass median diameters of several microns or greater [Duce, 1995; Savoie and Prospero, 1982], the phase function has a dominant forward scattering peak. Nonetheless, Mie calculation show that for "reasonable" size distributions (that is, a relatively broad size distribution comparable to that found in ambient aerosols) the backscatter fraction in the scattering angle range viewed by AVHRR is relatively insensitive to aerosol size over the range from 0.2-3.0 m diameter [R. Husar, unpublished data].

The AVHRR retreival algorithm has subsequently undergone a number of changes that will substantially improve the accuracy and precision or the resulting AOT values and may permit the retreival of additional aerosol parameters (see Stowe et al., this volume).

3.0 The AVHRR Aerosol Record: July 1989 - June 1991

This work presents a summary of the global oceanic aerosol distribution patterns as detected by the NOAA/11 polar-orbiting satellite between July 1989 and June 1991. The aerosol data after June, 1991 were greatly influenced by the Mt. Pinatubo volcanic eruption. During the pre-Pinatubo period the stratosphere was unusually clean [Dutton et al., 1994] and the backscattering signal was dominated by the tropospheric aerosol [Stowe et al., 1992; Long and Stowe, 1994; Stowe et al., this volume]. The results are presented in four seasonal (quarterly) maps (Fig. 1); in subsequent discussions we refer to the individual maps by the initials of the seasonal months (i.e., June-July-August as JJA).


Figure 1 (a-d) Radiatively equivalent aerosol optical thickness (EAOT x 1000) over the oceans derived from NOAA AVHRR satellites for the four seasons. The figure incorporates data for the period July 1989 to June 1991.

The seasonal EAOT data displayed in Figure 1 are available in electronic format by downloading in latitude, longitude grids in hdf files.

3.1 General Features of Aerosol Distribution Patterns

The EAOT distributions in Fig. 1 clearly show that the most prominent areas of increased EAOT are associated with continental sources. In many regions the continents are fringed by areas of high EAOT. In some regions the continents appear, in effect, to emit long "plumes" of increased EAOT. The continental aerosol plumes are generally characterized by high values of EAOT near the coasts, the values declining with distance from the coast. This aerosol pattern is consistent with the transport of aerosols from continental sources by large scale wind systems, followed by atmospheric dispersion and removal in the downwind direction. In well-defined flow fields such as the trade winds (e.g., the tropical NAO), the continental plumes are markedly elongated and can extend over several thousand kilometers.

A second type of EAOT distribution consists of isolated patches that do not appear to be linked to continental sources. These distributions show much weaker spatial gradients which would be consistent with more diffuse large-scale sources (e.g., oceanic emissions). An example is the belt of slightly increased EAOT values that extends over much of the southern oceans at 30o-60oS.

On the basis of the temporal and spatial distribution of EAOT in Fig. 1 we can identify ocean regions where the seasonal variability of EAOT displays a characteristic pattern; these regions are delineated in Fig. 2., using a rectangular geometry for computational convenience. The monthly averages and annual average EAOT for each region are presented in Table 1. The monthly average EAOT in each of these regions is shown in a series of graphs in Fig. 3a-p.

Figure 2 also shows the location of various types of potential aerosol sources on the continents; these provide a qualitative picture of the geographic distributions of emissions that might be important sources for the EAOT plumes depicted in Fig. 1. Most anthropogenic emissions are located in the northern hemisphere (NH) and they are concentrated in the mid-latitudes [Hameed and Dignon, 1992; Benkovitz et al., 1996; Husar and Husar, 1990], especially eastern North America, Europe, and eastern Asia. Figure 2 shows the distribution of sulfur sources which, in this context, is also broadly representative of other types of anthropogenic aerosol emissions and their precursors (e.g., organic species, nitrates, black carbon, etc.). Arid regions and deserts are potential sources of wind blown dust [Pye, 1987, 1989]. In Fig.2 we show the location of sand desert regions [Olson, 1986]. However, it should be noted that sand deserts are not necessarily good or exclusive sources of wind-blown dust, a point that will be emphasized in this paper.

Biomass burning is an important source of aerosols, especially black carbon [see Levine, 1991]. Emissions from the burning of savannas comprise about 50% of all tropical biomass burning sources [Hao and Liu, 1994]. To reflect the distribution of this type of source we show the distribution of tropical savanna [Olson, 1986]. The map also shows the locations where biomass burning was reported by NASA astronauts during space flights [Andreae, 1993].

Oceanic sources of atmospheric aerosols are also of interest. In the absence of pollutant aerosols, the dominant sub-micron aerosol component over the oceans is believed to be nss-SO4= that results from the oxidation of biologically-produced dimethylsulfide (DMS) emitted from the oceans [Shaw, 1987; Charlson et al., 1987; Covert et al., 1992]. Unfortunately, there are no satisfactory proxy indicators for oceanic sulfur emission patterns. For example, oceanic emissions of DMS do not appear to be related to chlorophyll distributions in the ocean [Bates et al., 1993; Covert et al., 1992] which can be measured by satellite.

Figure 1 shows that on a global scale the highest EAOT values occur during the summer season, and the lowest during the winter. The seasonal variability is more pronounced at mid latitudes than at low latitudes. The hemispheric seasonality is depicted in Figure 3a (see also Table 1) which shows the average optical depth for the NH, the SH, and the globe. (Note, however, that the hemispheric values do not include the polar regions which are not covered by the satellite data.) The global average shows only a small annual variability (0.10 < <0.14); individually, the NH and NH show a seasonal amplitude of about a factor of two. The NH average peaks at =0.20 in April-June, and the lowest values (=0.10) are in November and December. The SH has the highest monthly average in January (=0.13) and lowest in May (=0.07). The seasonal patterns in the two hemispheres are not quite symmetric: the NH peaks in late spring and early summer, while the SH is highest in mid summer.

The impact of industrial anthropogenic aerosol sources is most evident in the mid-latitude aerosol distributions between 30-60 in the NH and SH (Fig. 3p); to facilitate comparison, the plot of the SH monthly averages is shifted by six months to correspond seasonally to the NH. In the NH mid-latitudes the annual average oceanic EAOT (=0.12) is 1.5 times that of the SH (=0.08). It is interesting to note, however, that much of the increased EAOT in the NH occurs during the spring months (April-June) and it is attributable to aerosols over the North Pacific Ocean (NPO), as evidenced by the very large plume in Fig. 1/MAM.

In the following sections, we discuss the EAOT distributions in terms of our current knowledge of aerosols over the oceans. We compare the temporal and spatial variability of EAOT with aerosol chemical measurements obtained from various ocean regions. Because we are interested on seasonal time-scale variability, we refer principally to data sets that were acquired over relatively long time periods; this means that we must rely mainly on measurements made at the surface in the marine boundary layer. Satellite estimates of AOT integrate over the thickness of the atmospheric column whereas the aerosol measurements, at best, approximate the concentrations in the boundary layer. Nonetheless, over the time scale of the seasons, we might expect to see some correspondence between the ground-based and satellite data sets; indeed, one objective of this exercise is to see how well they agree.

We focus our discussions of aerosol data on those constituents that are most likely to have the greatest impact on the radiative properties of the atmosphere as viewed by AVHRR: non-sea-salt (nss) sulfate, mineral dust and biomass burning products. We do not include nitrate because previous studies have shown that over the ocean, most of the nitrate mass is associated with sea-salt aerosol in the supra-micron size range [Savoie and Prospero, 1982] where it constitutes only a small fraction of the aerosol mass. Thus the radiative effects of nitrate will be relatively small compared to other constituents [Li et al., 1996]. Similarly, ammonium, an important aerosol constituent, is concentrated in the submicron size range in association with nss-sulfate; as such it only constitutes a small fraction of the aerosol mass in the sub-micron fraction. We do not discuss sea-salt aerosols because they do not appear to have a sufficiently strong impact on atmospheric radiative processes to be readily detectable by AVHRR; note, for example the absence of any regions with substantially increased EAOT values duirng winter in either hemisphere in Fig 1.

In discussing the distribution and properties of aerosols in the context of the AVHRR data, we do not attempt to make a comprehensive review of all relevant aerosol literature; instead we only cite as examples those containing measurements made during the time of the AVHRR data shown in Fig. 1 or literature that contains relatively comprehensive or unique coverage.

In our assessment of the AVHRR product, we make extensive reference to Nimbus-7/TOMS data (Herman et al., this volume). On the basis of the relative radiance measured at 340nm and 380nm, Herman et al. obtain information on the global distribution of absorbing aerosols, largely mineral dust and black carbon derived from biomass burning (and, to a lesser extent, coal combustion). Because TOMS is not sensitive to sulfate aerosols and other pollution products, it does not "see", for example, the large pollution plumes in the mid-latitude North Atlantic Ocean (NAO). Thus, by comparing the AVHRR product with that from TOMS, we can distinguish among the various types of sources.

Before discussing the aerosol data, it is important to note that cloud contamination (that is, the misinterpretation of cloud as aerosol) does not appear to be a major problem. If the cloud screening algorithm introduced errors, we would expect to see, for example, high rates of anomalous EAOT from the regions bordering the intertropical convergence zone (ITCZ) and symmetrical with it. As can be seen in Fig. 1, aerosol plumes are often seen to the north of the climatological position of the ITCZ, but not to the south. This suggests that, in general, the EAOT "plumes" are not cloud-contamination artifacts. Clouds can bias the EAOT results presented here to the extent that the aerosol concentrations and properties during cloud free conditions are different from those during cloudy conditions and to the extent that the clouds reduce the data coverage from any specific region. Nonetheless, from the standpoint of direct radiative forcing, it is the aerosol distributions shown here, obtained under clear-air (i.e., cloud free) conditions, that are most relevant.

3.2 Atlantic Ocean

The quarterly composites (Fig. 1) clearly show that EAOT values for the NAO and the low-latitude SAO are relatively high during much of the year. In all cases the EAOT spatial distributions have a well-defined plume-like character that suggests that the aerosols are derived from continental sources. As shown in Fig. 2, there is a high density of pollutant sources in North America and Europe which together account for about half of the global total of anthropogenic sulfur and nitrogen emissions [Levy and Moxim 1989; Hameed and Dignon 1992]. Also, large quantities of soil dust are transported out of North Africa all year long, affecting much of the tropical NAO and Caribbean. Biomass burning is carried out extensively in Africa and South America (see Fig. 2) and combustion products are carried over large areas of the NAO and SAO in the low-latitudes. In the following sections we discuss the impacts of these sources on various regions in the Atlantic.

Tropical Atlantic Ocean and Caribbean. On a global scale, the largest and most persistent areas of high EAOT values are found over the tropical NAO. The plume is at its maximum extent in the summer when it reaches into the Caribbean, the Gulf of Mexico and the southeast coast of the United States. The relatively steep gradient on the southern boundary of the tropical NAO EAOT plume roughly corresponds to the seasonal climatological position of the northern boundary of the ITCZ. The plume shifts seasonally in a manner that is consistent with seasonal changes in the large scale circulation and the migration of the ITCZ towards the summer hemisphere.

The EAOT plume is clearly associated with the transport of dust from sources in North Africa. Large quantities of mineral dust are transported across the tropical Atlantic during much of the year [Carlson and Prospero, 1972; Prospero and Carlson, 1972; Karyampudi and Carlson, 1988]. Satellite imagery shows that it takes about one week for dust outbreaks to cross the tropical NAO from the coast of Africa to the Caribbean [Ott et al., 1991]. The main transport occurs at higher altitudes in a layer (the Saharan air layer) that typically reaches to several km and often to 5-6 km [Prospero and Carlson, 1972; Talbot et al., 1986; Karyampudi and Carlson, 1988; Westphal et al., 1987; 1988]. Concentrations aloft are usually several times greater than in the marine boundary layer. Because of the unusually high temperature and low relative humidity of the dust layer, it can be identified in routine meteorological soundings as far west as the Caribbean [Carlson and Prospero, 1972; Ott et al., 1991] and over the Amazon basin in eastern Brazil [Swap et al., 1992; Artaxo et al., 1994]. The placement and extent of the EAOT plume in JJA matches that obtained with sun photometers aboard a network of ships in the tropical NAO during the summer of 1974 [Prospero et. al., 1979].

The seasonal movement of the EAOT plume conforms precisely with the mineral dust measurements that have been made almost continuously at Barbados (13.17oN, 59.43oW) since 1965 [Prospero and Nees, 1986]. In Fig. 4a we show two years of Barbados daily dust concentration data that was acquired during the time corresponding to that of the satellite record. Dust concentrations typically rise and fall in a coherent manner over the period of several days with the passage of easterly waves [Carlson and Prospero, 1972; Prospero and Carlson, 1972; Karyampudi and Carlson, 1988] as substantiated by satellite imagery [Ott et al., 1991]. The monthly mean dust concentrations at Barbados (Fig. 5a) show a pronounced summer dust maximum with concentrations about ten times those in winter; the summer maximum at Barbados occurs when the core of the EAOT plume (Fig. 1, JJA) lies along the latitude of Barbados. The aerosol data are consistent with the seasonality shown in the Caribbean block (Fig. 3k) which also matches that of the West Africa block (Fig. 3b) although the seasonal amplitude is much greater in the former than the latter.

In addition to carrying large quantities of dust, the winds over the NAO often bring high concentrations of pollutants. The pulses of increased dust concentrations are accompanied by sharply increased concentrations of nss-SO4= (Fig. 4b) and NO3-. At Barbados, on an annual basis about half of the nss-SO4= is attributed to pollutants, most of which appear to be derived from sources in Europe [Savoie et al, 1989a; Savoie et al., 1992]. Biomass burning is widespread in tropical Africa [Delmas et al, 1991; Andreae, 1993; Brustet et al., 1991] during the NH late winter and spring, peaking in March-April [Hao and Liu, 1994]; biomass burning products are observed in aerosols over the tropical NAO [Andrea, 1983] and at Barbados [Savoie et al., 1992]. Nonetheless, on a mass basis, mineral dust is the major aerosol component in this region [Prospero, 1996a,b]. At Barbados during the past decade, the mean concentration of mineral dust is about ten times greater than the combined total of the other important non-sea-salt aerosol components (i.e., NO3-, nss-SO4=, and NH4+) [Prospero, 1996a; Li et al., 1996]. Aerosol light-scatter measurements on Barbados [Li et al., 1996] show that on an annual basis, light scatter by mineral dust is four times greater than that by nss-SO4=.

The seasonal cycle of aerosol concentrations in the Canary Islands is similar to that at Barbados and consistent with the AVHRR seasonal patterns. Studies made at an observatory at Izana, Tenerife (28.30şN, 16.50şW), at an altitude of 2360m, above the mean top of the marine boundary layer, provide data in the free troposphere within the Saharan air layer [Arimoto et al., 1995; Prospero et al., 1995a]. Tenerife is located on the northern edge of the main transport plume during much of the year (Fig. 1). As a result, aerosol concentrations are highly variable (Fig. 6a), depending on the day-to-day synoptic conditions. Although the aerosol concentrations on some days can be extremely high, the annual mean concentrations at this site are similar to those on Barbados [Prospero, 1996a,b]. The similarity in concentrations is largely due to the fact that Barbados lies in the path of the North African plume a larger fraction of the time than does Tenerife, in agreement with the depiction of the plume in Fig. 1. While dust transport to the Canary Islands can take place any time of year, depending on specific synoptic conditions, the frequency of events becomes much greater in mid summer when the islands lie on the north edge of the plume.

The EAOT values in the Caribbean and Central America blocks (Fig. 3k) show a maximum in the late spring and the summer. As suggested earlier, the seasonal cycle in the Caribbean is largely driven by the advection of African dust into the region. In contrast the Central American block values peak in the late spring because of emissions from regional sources, as suggested by the high EAOT values and steep gradients along the coast in this region (Fig. 1). Some of the Central American aerosol is probably transported from large urban regions (e.g., Mexico city). However, as discussed more fully below, there is also extensive biomass burning in this region in the spring.

The SE US block has a modest annual average EAOT=0.17. The monthly averages (Fig. 3l) show a very strong peak in spring and summer. The spring peak appears to be largely due to the impact of transport from Central American sources to the Gulf of Mexico region. The summer peak is due in part to the effects of North American pollutants; note the similarity in timing and peak values with that in the E US block (Fig. 3l, see below). In addition, the SE US block is strongly affected by the advection of African dust into the region as suggested by the similarity to the Caribbean block trend (Fig. 3k) and by dust measurements made in Miami (25.75oN, 80.25oW) [Prospero et al., 1987, 1993] and Bermuda (32.27oN, 64.87oW) [Arimoto et al., 1992, 1995], all of which show a very strong summer maximum.

During the NH winter the large scale circulation systems shift southward. In North Africa, winter is the season of the Harmattan [see Morales, 1979], when large quantities of dust are carried southward out of the dust source regions in the Sahara and the Sahel, producing dense hazes in the countries bordering the Gulf of Guinea [see Morales, 1979; Prospero, 1981]. Dust is lifted by convection to altitudes of several km and carried out over the Gulf of Guinea, above the southwesterly monsoon winds, consistent with the DJF EAOT distribution shown in Fig. 1. The EAOT is highest in February and March and lowest in September and October. In the Guinea block (Fig. 3c) the average EAOT during the peak months February and March is 0.48 - 0.49, a value comparable to the peak monthly EAOT's for West Africa (Fig. 3b) in June and July. However, as noted below, biomass burning products could contribute to these high values.

Thus, all evidence suggests that the AVHRR plume over the tropical NAO and Caribbean is largely attributable to African dust. This conclusion is supported by the TOMS data (Herman et al., this volume) wherein the plume has the same shape and dimensions and shows the same seasonal variability as that observed in AVHRR.

Central North Atlantic Ocean. Plume-like EAOT features are clearly evident over the mid-latitude NAO in Fig. 1. During the MAM and JJA, a large EAOT plume emerges from the middle-Atlantic states of the U.S. and extends to the central NAO, consistent with the prevailing westerly winds in this region. In spring the plume passes over Bermuda (32.27oN, 64.87oW). Aerosol nss-SO4= concentrations measured on Bermuda (Fig. 7a) are relatively high, reflecting the proximity of Bermuda to North American pollution sources [Galloway and Whelpdale, 1987; Arimoto et al., 1992; Arimoto et al., 1995; Ellis et al., 1993]. Tracer studies [D. L. Savoie, unpublished data] suggest that on an annual basis, 70% of the nss-SO4= aerosol at Bermuda is anthropogenic. Monthly mean nss-SO4= concentrations (Fig. 7b) are substantially higher in the spring when increased pollutant concentrations are transported to Bermuda behind fronts [Moody et al., 1995]. Concentrations can be relatively high in the summer as well (Fig. 7b) but transport from the west is much more variable because of the influence of the Bermuda-Azores high pressure center which frequently brings southerly winds to the region. This is consistent with Fig. 1, JJA, which shows the core of the EAOT plume passing somewhat to the north of Bermuda.

The aerosol measurements on Bermuda conform to aerosol trends implied by the seasonal EAOT pattern in the SE US block (Fig. 3l) where the monthly values peak in April-May (0.24-0.26) and in the E US block in June - July (0.24-0.25); the minimum is in the November - February. Note that the annual average EAOT in the E US block is rather modest, 0.15, compared to the W Africa block, 0.26, and the Caribbean, 0.20. The increased EAOT values during the summer (Fig. 3l) and the long plume over the NAO in JJA are consistent with seasonal trends in pollution emissions and haze [Husar and Wilson, 1993] and with AOT trends obtained from sun photometers in the eastern US [Husar et al., 1981; Trijonis et al., 1990]; recent studies of aerosol light extinction [Malm et al., 1994] show a summertime maximum in the region of the middle Atlantic states with over 50% of the extinction attributable to ammonium sulfate.

African dust is also a major aerosol constituent at Bermuda especially during the summer months in association with trajectories that come from the tropical NAO [Arimoto et al., 1995] and Africa, a feature that is consistent with the strong circulation associated with the Bermuda-Azores high. Dust transport to the SE coast of the US is suggested in the AVHRR/JJA EAOT distribution and it is clearly shown in the TOMS absorbing aerosol data [Herman et al., this volume]. At Bermuda there are dust events in the spring and fall but they appear to come from sources in North America; however the dust concentrations associated trajectories from the US are a factor of ten lower than those from Africa [Arimoto et al., 1995].

In the eastern NAO, the large area of increased EAOT in the spring and summer appears to be largely attributable to European pollution events, usually associated with the presence of a high pressure center over western Europe [Doddridge et al., 1994]. Pollutants are carried westward, around the high, and into the high latitudes. At Mace Head (53.32oN, 9.85oW), on the west coast of Ireland, nss-SO4= (and NO3-) aerosol concentrations peak sharply in the spring and summer (Fig. 8); 80-90% of the nss-SO4= is attributed to pollution sources [McArdle and Liss, 1995]. Measurements of aerosol size distributions and chemical characteristics at Mace Head [Jennings et al., 1991, O'Dowd et al., 1993] show that when the site is impacted by air masses from Europe the aerosol properties are typical of pollutant products; in contrast, the aerosols carried in westerly trajectories show minimal anthropogenic influences. Similarly, in the higher latitudes of the NAO, sharply increased concentrations of nss-SO4= and NO3- are observed at Hiemaey, Iceland (64.40oN, 20.30oW) due to pollutant transport events from Europe [Prospero et al., 1995b]; in the absence of such events, aerosol concentrations are usually quite low, comparable to values measured in the remote SH oceans.

During MAM and JJA, the North American plume merges with that from Europe, effectively bridging the NAO. The monthly mean EAOT values for the N. Atlantic block (Fig. 3n) are relatively constant (between 0.16 to 0.18) from April through August. The large area of relatively low EAOT values in the central NAO coincides with the mean position of the Bermuda-Azores high pressure center. On the basis of the quarterly composites, it is not possible to judge which of the two continental source regions is having the greatest impact on the central NAO. However, monthly composite images of the AVHRR EAOT data (not presented here) show a clear separation between these plumes. Based on the gradients in EAOT distributions, it would appear that transport from Europe is dominant during MAM. In contrast, during JJA, the North American sources seem to be more important. This interpretation is supported by the trend in monthly average EAOT values for the NW Europe and W Europe blocks (Fig. 3h); the annual cycles for both regions are essentially identical with a maximum in April-May (0.23-0.24) and a minimum in November-January (0.04-0.05).

Also during the spring and summer large quantities of pollutants are carried out of Europe southward across the Mediterranean [Bergametti et al., 1989a,b] and over the coastal waters off the Iberian peninsula and the North Africa. The average EAOT for the Mediterranean is 0.18. There is a strong seasonality in EAOT (Fig. 3i) with a summer maximum (August EAOT=0.29) and a winter minimum (December, 0.07). There is also evidence of a secondary peak in April that is comparable in timing and magnitude to those for the NW Europe and W Europe blocks (Fig. h). The seasonal composites (Fig. 1) show a gradient in EAOT values that increases toward the coast of North Africa; this gradient (which is more sharply defined in monthly AVHRR composites) is consistent with the observation [Bergametti et al., 1989a,b] that the transport of dust from North Africa to Europe is greatest during the summer and with the TOMS absorbing aerosol distributions [Herman et al., this volume].

Tropical and Subtropical South Atlantic. In DJF and MAM (Fig. 1) large amounts of North African dust are transported over the tropical NAO to the NE coast of South America [Prospero et al., 1981; Talbot et al., 1986]. At this time of year, African dust is an important aerosol constituent in the atmosphere over the Amazon basin where the dust is believed to serve as an important source of nutrients for the soils [Swap et al., 1992]. The relatively high EAOT values observed in the NE Brazil block (Fig. 3c) in January through March appear to be largely a consequence of this transport. EAOT values are relatively low during the later half of the year, reflecting the fact that this region lies in the SH circulation; nonetheless the EAOT values remain high (monthly means, 0.16-0.19). The AVHRR distributions suggest that this region is under the influence of transport from Africa all year long.

Biomass burning products are a major contributor to some of the low-latitude plumes seen over the SAO in Fig. 1. There is extensive burning in the savanna that covers much of the low latitudes in central and west Africa [Hao and Liu, 1994] as indicated by astronaut observations of fire (Fig. 2) and by TOMS [Herman et al., this volume]. The extensive EAOT distribution over the Gulf of Guinea during DJF is probably associated with both dust and biomass burning products. Modeling studies of CO2 transport from biomass fires [Iacobellis et al., 1994] shows a very prominent plume in January and February that fits very well with the DJF EAOT in Fig. 1.

In central Africa, savanna and grassland fires are most frequent in June while in southern Africa they occur mostly in July-September [Andreae et al., 1994]. The reported location and seasonality of the burning in southern Africa is consistent with the EAOT distributions shown in Fig. 1. The monthly average EAOT data for the SW Africa block show the maximum in August-September (0.29-0.32). Because of the position of the SAO high pressure center, in conjunction with meteorological processes over southern Africa, transport to the west is strongly favored during most of the burning season as shown in Fig. 1; transport to the east, into the Indian Ocean (IO), becomes significant in October but it is relatively minor compared to the transport over the SAO as indicated by the monthly EAOT data for the SE Africa block (Fig. 3d). During biomass burning studies in southern Africa in August to October 1992 [Andreae et al., 1994], it was noted that westward-moving smoke-laden air masses tended to exit over Angola, consistent with the EAOT distribution in Fig. 1; these parcels followed trajectories that crossed the coast of South America over central and northern Natal, Brazil, in agreement with the distribution in Fig. 1. These conclusions are all supported by TOMS [Herman et al., this volume].

During SON there is evidence of a weak equatorial plume that extends from the states along the Gulf of Guinea towards the NE coast of Brazil (Fig. 1). This plume appears to be distinct from the dust plume that emerges further to the north along the African coast and the prominent biomass-burning plume to the south, along the coast of Angola and Namibia. Aircraft measurements made in the low latitudes off the Brazilian coast during September 1989 [Andrea et al., 1994b] showed that haze layers were frequently present at altitudes of 1 to 5.2 km; the layers contained enhanced concentrations of aerosols and chemical constituents (e.g. O3, CO and oxides of nitrogen) that are characteristic of biomass burning products. These observations are consistent with the seasonal cycle of EAOT observed in the NE Brazil block (Fig. 3c).

Some of the EAOT in SON (Fig. 1) could be associated with dust transported out of arid regions in Angola and South Africa (e.g., the Namib and Kalahari deserts); however, the axis of the plume is somewhat further north than would be expected for dust from these regions.

The ocean region off the SE coast of Brazil is generally characterized by surprisingly low EAOT values (Fig. 3j) in light of the intense biomass burning that takes place throughout this region [Andreae, 1993; Cahoon et al., 1991; Artaxo et al., 1994]. In the Amazon basin, south of the equator, burning takes place during much of the dry season, July-September, but it is most intense in the late season, August-September [Artaxo et al., 1994; Hao and Liu, 1994; Herman et al., this volume]. Models [Iacobellis et al., 1994] show a major plume extending SE from central South America, the axis crossing the coast of Brazil between about 20o-30oS; a plume also extends to the NW, crossing the west coast of South America, roughly between 10oS to 10oN. Yet there is no evidence of any plume along the east coast in either JJA or SON in Fig. 1. The absence of strong evidence for biomass burning plumes is attributed to the fact that the export of smoke to the SE often appears to be associated with meteorological conditions that are accompanied by extensive cloud cover which would preclude detection by AVHRR. This scenario is consistent with TOMS [Hsu et al., 1996; Herman et al., this volume] which shows a marked peak in emissions in August-September; smoke from the northern Amazon basin is transported over the NW coast near Ecuador while smoke from the southern region is transported over the SE coast in association with cloudy conditions. Nonetheless, in agreement with AVHRR, the TOMS data [Herman et al., this volume] suggest that the export of biomass burning products from South America is much less extensive than from south Africa.

3.3 Indian Ocean, Arabian Sea and the Bay of Bengal

The annual mean EAOT for the Arabian Sea is 0.32, the highest of all regions in Table 1. EAOT in the Arabian Sea block (Fig. 3b) peaks sharply in JJA (0.61-0.65); the seasonal cycle is similar to that for the West Africa block, suggesting that the entire region is dominated by similar dust-forcing meteorological and climatological processes. Soils in the Tigris and Euphrates basin appear to be a major source of dust that is transported to the Arabian Sea [Ackerman and Cox, 1989; Prospero, 1981; Khalaf et al., 1985] although soils in the arid regions of NW India and East Africa could also contribute [Ackerman and Cox, 1989]. The seasonality of dust storms and haze conditions in this region [Ackerman and Cox, 1989] is consistent with the seasonality of the EAOT distributions in Fig. 1. During JJA, much of the transport takes place in deep atmospheric layers that extend to 4-7 km and have properties similar to those observed with Saharan dust outbreaks [Ackerman and Cox, 1989]. It is important to note that the SW summer monsoon is well established over the Arabian Sea in June; the fact that high values of EAOT in JJA (Fig. 1) extend so far to the south is attributed to the transport of dust over the top of the monsoon inversion [Ackerman and Cox, 1989; Sirocko and Sarnthein, 1989]. In the low latitudes and at low altitudes, the SW monsoon winds carry dust from the Horn of Africa into the Arabian Sea [Sirocko and Sarnthein, 1989].

Although dust storm statistics suggest that there are major sources of mineral dust in northern India [Ackerman and Cox, 1989], there is very little dust storm activity in southern India below 15oN at any time of year. Thus the prominent bulge in the EAOT distribution (Fig.1) located off the SW coast of India in MAM is most likely due to pollution aerosol, not dust.

EAOT values over the Bay of Bengal are substantially lower than over the Arabian Sea (annual mean, 0.22). EAOT values are highest in JJA, similar to the Arabian Sea, with the maximum monthly average in June (Fig. 3f, 0.39). The gradients in Fig. 1 suggest that the sources lie mostly in India , but transport of dust across the subcontinent from the Arabian Sea is a possibility. In addition, monthly EAOT composites (not shown here) suggest that there are substantial sources in Bangladesh and northern Burma; although biomass burning is common in these regions, the timing of the enhanced EAOT does not match the peak burning period, March through May [Hao and Liu, 1994], which is also shown in TOMS [Herman et al., this volume].

There is very little aerosol data for the IO with which to compare the AVHRR distributions; the data that is available has been largely obtained over relatively short periods during ship cruises. Nonetheless, these clearly show that dust concentrations are very great over the Arabian Sea and the NW IO close to Africa, comparable to those measured along the west coast of Africa and the Mediterranean [Savoie et al., 1987; Prodi et al., 1983]. These values and the seasonality of the concentrations are generally consistent with the dust transport distributions shown in AVHRR (Fig. 1) and with the monsoon circulation [Ackerman and Cox, 1989]. Although dust sources appear to be dominant in the Arabian Sea and possibly the Bay of Bengal, substantial concentrations of pollutant species are also present [Savoie et al., 1987].

The EAOT distributions over the Arabian Sea in JJA show a very strong gradient in the low latitudes, close to the equator, in the vicinity of the ITCZ . Dust concentrations drop sharply as one moves from the Arabian Sea southward into the SH circulation [Savoie et al., 1987; Prodi et al., 1983]. On a recent cruise in the southern and central IO [Dickerson et al., 1996], nss-SO4= concentrations south of the ITCZ were generally in the range 0.1-0.5 µg/m3, values that are typical of background ocean values; north of the ITCZ, concentrations were ten times greater with values as high as 9 µg/m3. The concentration of CO also showed large temporal variability that was highly correlated with aerosols, which suggests that both species were derived from pollution sources on the Asian subcontinent.

Although the ITCZ represents a demarcation line in aerosol distributions, the EAOT gradients in JJA suggest that substantial amounts of aerosol material are being transported south wind, perhaps as far as 15o-20oS. During the remainder of the year, the eastern tropical IO appears to be affected by other aerosol sources in Oceanea, Australia and southern Africa. The monthly mean EAOT values for the Indonesia block (Fig. 3e) have a strong seasonal pattern with peak values in September-October, possibly due to extensive biomass burning in this region (see Fig. 2) [Malingreau et al., 1985]. This period corresponds to the end of the dry season; the large scale winds would transport burning products to the west, over the IO. The EAOT map for SON (and to a lesser extent, DJF) shows a coherent aerosol plume that appears to originate from these islands and northern Australia and that extends over the IO to East Africa. The New Guinea block (Fig. 3e) does not reveal a seasonal trend but EAOT values are moderately high all year long. The aerosol emissions in this block would include materials from northern Australia which has a long dry season (from April-May through the end of the year), and from the eastern portion of Indonesia; again, we might assume that biomass burning could be a major source based on the frequent sightings as indicated in Fig. 2. However the TOMS data [Herman et al., this volume] do not show a great deal of burning in this region; the burning that is observed takes place in September-November, consistent with the trend of EAOT. Six years of atmospheric turbidity measurements at Broome on the NW coast of Australia (17.97oS and 122.23oE) also show a strong seasonal cycle with a maximum in September-November and a minimum in April-June [Scott et al., 1992].

The EAOT distributions do not reveal any persistent, substantial, well-defined plumes emerging from Australia. The absence of dust plumes is particularly noteworthy considering that Australia has the largest expanse of desert land in the SH [Pye, 1987].

EAOT values over the southern IO are quite low most of the year (annual average, 0.08). Nonetheless, there is a pronounced seasonal cycle (Fig. 3m) with a minimum in June-July (0.04) and a maximum in January (0.13). The summer maximum is associated with a clearly visible band of enhanced EAOT values (Fig. 1) that circles the southern oceans between 40o-60oS in DJF. This will be discussed in a later section.

3.4 Pacific Ocean

From Fig. 1 it is clear that the NPO is much more heavily impacted by aerosols than the SPO. This difference is reflected in the annual mean EAOT's for the NE Pacific and EC Pacific blocks (0.11 and 0.13, respectively) compared to the SE Pacific (0.07) and in the very different seasonal cycles in the monthly EAOT's (Fig. 3m, 3n, 3o). In the following sections we discuss the data for various regions in the NPO and SPO.

North Pacific. In Fig. 1 in MAM a plume extends from Asia across the central NPO, almost to the west coast of Alaska and Canada. High EAOT values are visible along the Asian coast in all seasons. In contrast, the west coast of North America shows relatively low EAOT values at all times. Because of the presence of the high pressure center in the eastern NPO, this region is climatologically dominated by strong westerly and northerly flow in all seasons; consequently, transport of aerosols from the western US is not favored. Thus, the aerosol distributions over the mid-latitude NPO seem to be largely derived from sources in Asia.

There is a strong seasonal cycle which is most pronounced in the monthly EAOT averages in the NW and NE Pacific blocks (Fig. 3n) but it is also well-defined in the EC and WC Pacific (Fig. 3o). The seasonal cycle is largely driven by the transport of aerosols out of Asia as suggested by the plume in Fig. 1 and by the close match between the seasonal cycle for the NW Pacific block with those for the Japan (Sea of Japan) and China (Yellow Sea) blocks (Fig. 3g). The high coastal EAOT values are associated with mineral dust and pollutant aerosols [Arimoto et al., 1996; Mukai et al., 1990; Tsunogai et al., 1985; Uematsu et al., 1992; Prospero, 1996a; Gao et al., 1996]. The seasonality of the transport in this region is consistent with the large scale climatology: storms and cold outbreaks during the winter and spring favor transport out of Asia; the southerly-southeasterly monsoons bring relatively clean ocean air into the coastal regions in the mid and low latitudes beginning in late spring and reaching their maximum northerly extension in July.

The dust maximum corresponds to the seasonal cycle of dust storm activity in Asia [Goudie and Middleton, 1992; Littmann, 1991; Middleton, 1991; Merrill, 1989; Prospero et al., 1989; Duce, 1995]. Also, in the spring, strong westerly flow in the mid latitudes favors the transport of dust over great distances [Merrill, 1989; Duce, 1995; Prospero et al., 1989; Gao et al., 1992]. In Japan and Korea during the spring, they frequently experience extensive hazes that are caused by yellow dust (Kosa) that can be traced to sources in Asia [Takayama and Takashima, 1986; Tsunogai et al., 1985; Mukai, 1990; Uematsu et al., 1992]. Asia is also a major source of anthropogenic sulfur and nitrogen emissions [Hameed and Dignon, 1992; Husar and Husar, 1990; Galloway et al., 1994; Kasibhatla et al., 1993] and biomass burning products [Cahoon et al., 1991]. Coal is a major energy source and emissions of black carbon appear to be very high making these sources visible in TOMS in the winter months [Herman et al., this volume].

Aerosol measurements made in a network of surface-based stations in the NPO substantiate the seasonal character of the aerosol distribution in Fig. 1. In MAM the concentrations of soil dust, nss-SO4=, and NO3- increase sharply across a large area of the central NPO [Prospero et al., 1989; Savoie et al., 1989b; Arimoto et al., 1996]. At Midway (28.22oN, 177.35oW) which lies at the southern edge of the plume in MAM (Fig. 1), there is a strong spring maximum in dust concentration (Fig. 9a). Nss-SO4= also shows a pronounced maximum in the spring (Fig. 9b), along with the dust, suggesting a common source region; a similar spring maximum occurs for NO3- aerosol [Savoie et al., 1989b]. Tracer studies [Savoie et al., 1989b] suggest that during the spring a major fraction of nss-SO4= is derived from anthropogenic sources. Enhanced concentrations of these aerosol species are also observed at Shemya, in the Aleutians (52.92oN, 176oE) and on Oahu, Hawaii (21.33oN, 157.70oW) [Prospero et al., 1989; Savoie et al., 1989b; Arimoto et al., 1996]. In MAM, sharply increased concentrations of mineral dust [Holmes and Zoller, 1996] and NO3- [Lee et al., 1994] are also observed in the free troposphere at Mauna Loa Observatory (19.53oN, 155.57oW) on the island of Hawaii; the advection of Asian dust clouds into the region is also readily apparent in measurements of solar spectral irradiance and atmospheric transmission at Mauna Loa [Dutton et al., 1994].

Equatorial Pacific. major oceanic aerosol belt stretches over the Pacific just north of the equator (0o-20oN) in MAM (Fig. 1) along the northern edge of the ITCZ. (The belt is also present, albeit more weakly, in JJA.) The uniform aerosol characteristics across this huge region in MAM are reflected in the similarity of both the magnitude and the seasonalality of the EAOT values in the EC (Eastern Central) Pacific and WC (Western Central) Pacific blocks (Fig. 3o). While this region is occasionally impacted by the transport of Asian aerosol in the spring as discussed above, the plume, which lies in a band of easterly trade winds, appears to emanate from Central America between 10o-15oN. This plume is different from the other plumes discussed thus far in that it shows no gradient "down-wind"; indeed, the plume appears to broaden and intensify over the western NPO.

There are no well-quantified sources in Mexico and Central America that could explain such a prominent plume. Biomass burning is quite intense throughout this region (Fig. 2), mostly in April-May [Hao and Liu, 1994]. In April-May TOMS [Herman et al., this volume] shows very extensive areas of absorbing aerosol in precisely the same region where EAOT values are high in Fig. 1; the TOMS data would seem to implicate biomass burning sources but TOMS does not show any extensive plumes comparable to those from burning in south Africa. Nonetheless, it is difficult to see how emissions from sources in the boundary layer could be transported such a great distance over the ocean unless there is an efficient mechanism to lift the emissions into the free troposphere. Mexico City, at an elevation of 2,303 m, might be an effective source for injecting large quantities of pollutants into the free troposphere but it would seem unlikely to be the source of such a long plume. There was a major volcanic eruption in Guatemala (Pacaya, volcanic explosivity index of 3*) in January 1990 and it continued through 1993 [Simpkin and Siebert, 1995]. Perusal of the original weekly AVHRR product reveals that the plume is visible in both 1991 and 1990 although it is more prominent in 1991. However the fact that the EAOT plume in Fig. 1 has a distinct seasonality would seem to preclude a volcanic source.

The EAOT plume is not explained by cirrus cloud contamination. Wylie, et al. [1994] analyzed thin cirrus with the NOAA/HIRS CO2 slicing technique over global oceans from 1989 to 1993 and concluded that there is a high frequency of cirrus (greater than 50%) in the ITCZ in all seasons; also there is a modest seasonal movement that tracks the sun. Thus, if there was an artifact caused by thin cirrus, we would expect to see it in all seasons, not just in the MAM period.

There is very little aerosol data from the equatorial Pacific. Measurements made continuously over the period 1981 to 1986 [Savoie et al., 1989b] in the easterly trade winds at Fanning Island (3.92oN, 159.33oW) yielded moderately high nss SO4= concentrations (annual mean, 0.76 µg/m3); however, concentrations are relatively uniform all year long with monthly means falling in a rather narrow range of 0.60-0.74 µg/m3 [Savoie et al., 1989b]. Although transport from Asian sources is occasionally noted at Fanning in the spring, the impact of continental sources is believed to be small and sporadic. The absence of strong continental impacts at Fanning suggests that the nss-SO4= is largely derived from oceanic sources of DMS, a conclusion that is supported by the measured concentration of methanesulfonate (MSA) an oxidation product of DMS [Savoie et al., 1989b]. As was the case with nss-SO4=, MSA shows no seasonal variability, an observation that is consistent with the absence of any strong seasonality in the oceanic productivity in this region. Furthermore, extensive measurements along 140oW in the spring of 1992 found that oceanic DMS concentrations were highest between 3oN to 15oS [Huebert et al., 1994; Bates et al., 1993], considerably south of the EAOT plume which is located roughly between 8o and 16oN (Fig. 1); the DMS concentrations south of the equator were as much as twice those to the north. Yet there is no evidence of any comparably prominent oceanic-nss-SO4= plume to the south of the equator in this season or in any other season. Thus, the EAOT plume observed in this region in the spring does not seem to be directly related to oceanic sources of nss-SO4=.

The absence of obvious impacts form oceanic DMS sources is emphasized by that fact that there is no sign of any prominent source of enhanced EAOT values in the region off the NW coast of South America where there is very strong upwelling, making these waters among the most productive in the world. There is a narrow belt of increased EAOT along the coast, adjacent to Peru (see the Peru block, Fig. 3j), but this appears to be attributable to sources in South America, including biomass burning (Fig. 2). Sources in Brazil and Argentina should not have a major impact on this ocean region; transport to the west is hindered by the Andes. TOMS [Herman et al., this volume] does not show major transport of burning products to this region. Also, in the eastern SPO near the coast of South America there is a strong semi-permanent high pressure center at about 25o-30oS which results in southerly winds that tend to lie parallel to the coast all year long; the presence of the high is reflected in the position of the low EAOT values in this region in Fig. 1. Pollution sources could be significant here; high concentrations of nss-SO4= have been measured in the coastal ocean regions in the Peru block; these are largely attributed to smelters in Chile [Saltzman et al., 1986].

Mid-latitude South Pacific. In general, EAOT values throughout this region are quite low with both the SE Pacific and New Zealand blocks (Fig. 3m) having annual means of 0.07. Furthermore there is a strong seasonal cycle in the New Zealand block that is identical to that for the SE Pacific and S. Indian Ocean blocks. This suggests that sources in New Zealand and Australia have no discernible effect on EAOT in the western SPO. Extended measurements made on Norfolk Island (29.08oS, 167.98oE) and New Caledonia (22.15oS, 167.00oE) [Prospero et al., 1989] and on New Zealand [Arimoto et al., 1990] confirm that aerosol concentrations, including mineral dust, are generally quite low throughout this region.

The most prominent feature in Fig. 1 is a band of enhanced EAOT that circles the southern oceans between 45o-55oS during DJF, the SH summer. It is interesting that statistics on the distribution of haze at sea compiled from ship meteorological observations prior to the 1930's [MacDonald, 1938; see also Prospero, 1981] show a band of increased haze in precisely the same latitudes as that shown by AVHRR in DJF (but not in other seasons). The similar EAOT signatures across this huge area suggests that the AOT in this region might be due to natural oceanic sources. As previously stated, it seems unlikely that this feature could be due to sea-salt aerosols. Sea salt aerosol concentrations should be at a maximum in the winter when wind velocities are highest. A global model of sea-salt aerosol distributions (Erickson et al., 1986) shows a well-defined band of enhanced sea-salt aerosol concentrations in precisely the same latitude band as the band of enhanced EAOT in Fig. 1 in DJF; however, on an annual basis, the model shows substantially higher aerosol concentrations in DJF relative to JA.

It might be possible to ascribe the enhanced EAOT values in this band to oceanic nss-SO4= sources. The waters in the low-latitude southern oceans are highly productive at this time of year and the atmospheric concentrations of MSA are at a maximum [Mihalopoulus et al., 1993]. Measurements made at Cape Grim (40.68oS, 144.68oE) show a clear annual cycle of MSA and nss-SO4= concentrations with peak values in DJF [Ayers et al., 1991; Gras, 1995] yielding a seasonal average of about 0.3 µg/m3 for nss-SO4=. Similarly, measurements at Mawson (67.60oS, 62.50oE) on the Antarctic coast in the IO and Palmer Station (64.77oS, 64.05oW) show an extremely strong seasonal cycle in both MSA and nss-SO4= aerosol concentrations that peak at this time of year [Savoie et al., 1993]; however, even during spring, the mean nss-SO4= concentration is only a little over 0.2 µg/m3. Thus, the mean aerosol nss-SO4= values at remote southern ocean locations are typically about a factor of ten lower than concentrations in the NAO during the seasons when the EAOT plumes are clearly visible. Thus we would expect that the EAOT values due to ocean nss-SO4= would be rather low and perhaps not readily detectable in the present generation of AVHRR satellites.

4.0 Conclusions

The seasonal distributions of AVHRR EAOT over the oceans are clearly consistent with the long-term data sets for major aerosol constituents in the marine boundary. Similarly, the distribution of the EAOT plumes and their seasonal movement are generally consistent with our understanding of the meteorology and climatology in the source regions and the large scale circulation systems that transport the aerosol products. Aerosol measurements, although limited, provide a semiquantitative indication of the species that are most important in aerosol light scattering processes over the oceans. These studies clearly show the importance of anthropogenic products such as sulfate aerosol in light scattering processes. These impacts are most clearly evident over the mid latitude NAO and NPO where the produce long "plumes"; also, areas with very high EAOT values are often noted along the coasts, most prominently in Asia.

Nonetheless, the most notable features in the AVHRR images are those that are attributable to mineral dust and biomass burning products. The impact of dust is especially evident. The EAOT plumes attributed to dust yield the highest EAOT values and cover that largest areas. Dust plumes are most prominent over the tropical NAO and the Arabian Sea, reflecting the impact of sources in North Africa and the Middle East; AVHRR suggests that, from the standpoint of long range transport, these regions are the largest dust sources in the world, a conclusion that is substantiated by TOMS [Herman et al., this volume]. It would appear that the activity of the dust sources is somehow linked to the strong, deep and persistent convective mixing that takes place over the deserts and arid regions of North Africa [Karyampudi and Carlson, 1988; Westphal et al., 1987; 1988]. Once the dust is lifted high above the surface, it can be rapidly transported long distances by the vigorous wind systems that are found in the middle troposphere. Thus, the strong dust transport that we observe over North Africa and the Middle East is perhaps unique because of the convergence of these several factors.

While the AVHRR images clearly show that dust is a prominent feature, they also suggest that deserts are not necessarily good sources of dust. As previously noted, there is no visible dust plume associated with Australia, a continent that is largely arid. Furthermore, with regard to African sources, the very large dust plumes in DJF and MAM are seen to emerge from the sub-Saharan (Sahel) region; the Saharan sources appear to be dominant only in JJA and SON. It is believed that land use practices in the Sahel, coupled with drought, have greatly increased the rates of wind erosion [Williams and Balling, Jr., 1996]. This is supported by the observation of greatly increased dust transport across the NAO since the onset of drought in the sub-Sahara in the late 1960's [Prospero and Nees, 1986; Prospero et al., 1993]. Land use is implicated as an important factor in soil deflation, [Williams and Balling, Jr., 1996]; under such circumstances, dust might properly be regarded as an anthropogenic pollutant. The complexity of the dust deflation processes and the importance of land disturbance [Pye, 1989; Williams and Balling, Jr., 1996] is reflected in the difficulty that models have in replicating large-scale dust transport distributions compared to AVHRR [Tegen and Fung, 1994, 1995].

Finally, there is mounting evidence that dust can play an important role in global climate forcing. Early work had shown that dust has a substantial affect on the radiation balance in dust outbreaks over the NAO [Carlson and Benjamin, 1980] and that this forcing plays a role in the large-scale dynamics of the outbreaks [Karyampudi and Carlson, 1988]. More recently, models have shown that these effects have important implications for climate [Lacis and Mishchenko, 1995; Andreae, 1995, 1996; Tegen et al., 1996]. Thus, AVHRR can play an important role in characterizing the temporal and spatial variability of dust transport on a yearly basis and long term trends that might result from changes in climate and land-use.

Biomass burning is also an important source of EAOT plumes. However burning in southern Africa is the only source that seems to produce a persistent, large, clearly identifiable plume over the ocean. Other important burning sources are inferred (e.g. South America) but their impact seems to be much more limited and subtle. It may be that in many regions the smoke from fires is not lofted into the free troposphere where it can be transported over great distances. In general, AVHRR suggests that the areal coverage and EAOT resulting from burning is considerably less than that from mineral dust, a conclusion supported by TOMS [Herman et al., this volume].

Aerosols from conventional pollution sources (e.g., energy consumption), clearly have a great impact in the northern hemisphere EAOT distributions, especially in the mid latitudes. However, as previously stated, the mean EAOT values for these regions are considerably less than those affected by mineral dust and biomass burning sources. Nonetheless, it must be remembered that the current AVHRR algorithm may be discriminating against aerosols present at concentrations that yield low EAOT values; thus, the effects of conventional pollution aerosols (and other types of aerosols) could be more widespread (albeit at low concentrations) than indicated in Fig. 1.

The AVHRR images do not show any evidence of strong oceanic sources of aerosol. The prominent band of high EAOT values in the southern oceans might be attributed to oceanic sources of nss-SO4= aerosol. However, it is difficult to speculate on DMS-derived sulfate distributions over the oceans because of the considerable uncertainties about the relationship of DMS emissions to primary productivity (especially to chlorophyll distributions either directly measured or based on satellite-sensed values) and the subsequent conversion rates to sulfate aerosol. These uncertainties are reflected in the widely varying distributions obtained in global models of the atmospheric sulfur cycle. For example, Chin et al. [1996 - in press, JGR] show a distinctive band of increased concentration of DMS, MSA and nss-SO4= in January in the southern oceans, but the band is centered at about 60o S, whereas the band in AVHRR is centered at about 50oS; furthermore, there is little overlap between the model distribution and that viewed by satellite. In contrast, Pham et al. [1995] show no distinctive band of enhanced DMS in the mid- and high latitude southern ocean; their model yields the highest DMS emissions in equatorial regions. There is considerable research on role of oceanic sources of aerosols [e.g., Huebert et al., 1994]; the AVHRR data suggests that the region between 40o-60oS area is a prime candidate for study.

Another puzzling feature is the large plume that extends over much of the equatorial NPO in MAM. We speculated on possible sources in Central America (i.e., biomass burning and pollutants); but it is difficult to see how these sources could affect such a large region of the ocean unless they were transported in the middle and upper troposphere. Alternatively, we considered the possible role of ocean sources of nss-SO4=; however, existing data seems to argue against this source.

Taken as a whole, the AVHRR data show in a distinctive way that the NH is much more heavily impacted by continental sources than the SH. This is consistent with our understanding about the global distribution of pollution sources. It is also notable that there is a pronounced seasonality aerosol transport, especially in the mid latitudes; in general the largest plumes and higher EAOT levels are seen in the spring and summer hemispheres. This is true for conventional pollutants (e.g., sulfates) and also for dust and burning, an observation that is consistent with the TOMS product [Herman et al., this volume].

In order to asses the role of aerosols in climate, models require detailed knowledge about aerosol properties [Penner et al., 1994; Lacis and Mishchenko, 1995]. AVHRR provides a global-scale picture of aerosol distributions that can be used to guide the design and implementation of intensive aerosol campaign-type measurements and the development of both diagnostic and prognostic models of aerosol sources, transformations, deposition, and effects. Focused measurements at a few regionally representative sites would suffice to characterize the important aerosol properties [Stowe et al., 1990]. Future integrated earth observing satellite systems [King et al., 1992] will be capable of more detailed, long-term monitoring of global aerosol changes and their gross physical properties [Kaufman, 1995] as well as their relationship to climatic and other bio-geochemical variables and processes. Nonetheless, it will still be necessary to obtain detailed in situ aerosol and radiation data to further improve satellite aerosol retrieval algorithms and to validate the data that is subsequently obtained. The satellite data could then be used to extrapolate the in situ observations to larger space and time scales.

Acknowledgments. We thank Craig Long for providing the aerosol data files and Attila P. Husar for preparing contour maps. This research was partially supported by the NOAA grant #NA 16RC0517 and the National Science Foundation's Atmosphere/Ocean Chemistry Experiment (AEROCE) grants - ATM9414808, ATM9414812, and ATM9414846.

5.0 References

Ackerman, S. A., and S. K. Cox, Surface weather observations of atmospheric dust over the southeast summer monsoon region, Meteorol. Atmos. Phys., 41, 19-349 1989.

Andreae, M. O., Global distribution of fires seen from space, EOS, Trans., Amer. Geophys. Union, 74, 129-135, 1993.

Andreae, M. O., Climatic effects of changing atmospheric aerosol levels, in World Survey of Climatology, Vol. 16, Future Climates of the World, edited by A.. Henderson-Sellers, pp. 341-392, Elsevier, Amsterdam, 1995.

Andreae, M. O., Raising dust in the greenhouse, Nature, 380, 389-390, 1996.

Andreae, M. O., J. Fishman, M. Garstang, J. G. Goldammer, C. O. Justice, J. S. Levine, R. J. Scholes, B. J. Stocks, A. M. Thompson, B. van Wilgen, and the STARE/TRACE-A SAFARI-92 Science Team, Biomass burning in the global environment: First results from the IGAC-BIBEX field campaign STARE/TRACE-A/SAFARI-92, in Global Atmospheric-Biospheric Chemistry, edited by R. G. Prinn, pp. 83-101, Plenum Press, New York, 1994.

Arimoto, R., B. J. Ray , R. A. Duce, A. D. Hewitt, R. Boldi, and A. Hudson. Concentrations, sources, and fluxes of trace elements in the atmosphere of New Zealand, J. Geophys. Res., 95, 22, 389-22,405, 1990.

Arimoto, R., R. A. Duce, D. L. Savoie, and J. M. Prospero, Trace elements in aerosol particles from Bermuda and Barbados: Concentrations, sources and relationships to aerosol sulfate, J. Atmos. Chem., 14, 439-457, 1992.

Arimoto, R., R. A. Duce, B. J. Ray, W. G. Ellis, Jr., J. D. Cullen, and J. T. Merrill, Trace elements in the atmosphere over the North Atlantic, J. Geophys. Res. 100 (D1), 1199-1213, 1995.

Arimoto, R., R. A. Duce, D. L. Savoie, J. M. Prospero, R. Talbot, J. D. Cullen, U. Tomza, N. F. Lewis, and B. J. Ray, Relationships among aerosol constituents from Asia and the North Pacific during PEM-West A, J. Geophys. Res., 101 (D1), 2011-2024, 1996.

Artaxo, P., F. Gerab, M. A. Yamasoe, and J. V. Martins, Fine mode aerosol composition at three long-term atmospheric monitoring sites in the Amazon Basin, J. Geophys. Res. 99, 22,857-22,868, 1994.

Ayers, G. P., J. P. Ivey, and R. W. Gillett, Coherence between seasonal cycles of dimethyl sulphide, methanesulphonate and sulphate in marine air, Nature 349, 404-406, 1991.

Bates, T. S., K. C. Kelly, and J. E. Johnson, Concentrations and fluxes of dissolved biogenic gases (DMS, CH4 , CO, C02) in the equatorial Pacific during the SAGA 3 experiment, J. Geophys. Res., 98(D9), 16,969-16,978, 1993.

Benkovitz, C. M.., J. Dignon, T. Scholtz, J. Pacyna, L. Tarrason, J. Dignon, E. C. Voldner, P. A. Spiro, J. A. Logan, and T. E. Graedel, Global gridded inventories of anthropogenic emissions of sulfur and nitrogen, J. Geophys. Res., submitted, 1996.

Bergametti, G., A.-L. Dutot, P. Buat-Menard, R. Losno, and E. Remoudaki, Seasonal variability of the elemental composition of atmospheric aerosol particles over the northwestern Mediterranean, Tellus, 41B, 353-361, 1989a.

Bergametti, G., L. Gomes, E. Remoudaki, M. Desbois, D. Martin, and P. Buat-Menard, Present transport and deposition patterns of African dusts to the North Westem Mediterranean, in Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmosphetic Transport, pp. 227-252, Kluwer Academic Pubs., Norwell, Mass., 1989b.

Brustet, J. M., J. B. Vickos, J. Fontan, K. Manissadjan, A. Podaire, F. Lavenu, Remote sensing of biomass burning in West Africa with NOAA-AVHRR in Global Biomass Buming, edited by J.S. Levine, pp. 47-52, The MIT Press, Cambridge, London, 1991.

Cahoon, D. R., J. S Levine, W. R. Cofer, III, J. E. Miller, P. Minnis, G. M. Tennille, T. W. Yip, B. J. Stocks, P. W. Heck, The great Chinese fire of 1987; A view from space, in Global Biomass Buming, edited by J.S. Levine, pp. 61-66, The MIT Press, Cambridge, London, 1991.

Carlson, T. N., and S. G. Benjamin, Radiative heating rates for Saharan dust, J. Atmos. Sci., 37:193­213, 1980.

Carlson, T. N., and J. M. Prospero, The large-scale movement of Saharan air outbreaks over the northern equatorial Atlantic, J. Appl. Meteorol., 11, 283-297, 1972.

Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren, Oceanic phytoplankton, atmospheric sulphur, cloud albedo, and climate, Nature, 326, 655-661 1987.

Charlson, R. J., S. E. Schwartz, J. M. Hales, R. D. Cess, J. A. Coakley, Jr., J. E. Hansen, and D. J. Hofmann, Climate forcing by anthropogenic aerosols, Science, 255, 423-430, 1992.

Chin, M., D. J. Jacob, G. M. Gardner, M. Foreman-Fowler, P. A. Spiro, and D. S. Savoie, A global three-dimensional model of tropospheric sulfate, J. Geophys. Res, accepted May, 1996.

Covert, D. S., V. N. Kapustin, P. K. Quinn, and T. S. Bates, New particle formation in the marine boundary layer, J. Geophys. Res., 97(Dl8), 20,581-20,589, 1992.

Delmas, R. A., P. Loudjani, A. Podaire, and J­C. Menaut, Biomass burning in Africa: An assessment of annually burned biomass, in Global Biomass Burning: Atmospheric, Climatic, and Biospheric Implications, edited by J. S. Levine, pp. 126­132, MIT Press, Cambridge, Mass., 1991.

Doddridge, B. G., P. A. Dirmeyer, J. T. Merrill, S. J. Oltmans, and R. R. Dickerson, Interannual variability over the eastern North Atlantic Ocean: Chemical and meteorological evidence for tropical influence on regional­scale transport in the extratropics, J. Geophys. Res., 99, 22,923­22,948,1994.

Duce, R. A., Sources, distributions, and fluxes of mineral aerosols and their relationship to climate, in, Dahlem Workshop on Aerosol Forcing of Climate (1994, Berlin), edited by R. J. Charlson and J. Heintzenberg, pp. 43­72, Berlin, 1995.

Durkee P. A., D. R. Jensen, E. E. Hindman, T. H. Vonder Haar, The relationship between marine aerosols and satellite detected radiance, J. Geophys. Res., 91, 4063-4072, 1986.

Durkee, P. A., F. Pfeil, E. Frost., and R. Shema, Global analysis of aerosol particle characteristics, Atmos. Environ., 25A,. 2457-2471, 1991.

Dutton, E. G., P. Reddy, S. Ryan, and J. J. DeLuisi, Features and effects of aerosol optical depth observed at Mauna Loa, Hawaii: 1982-1992, J. Geophys. Res., 99, 8295-8306, 1994.

Erickson, D. J. III, J. T. Merrill, and R. A. Duce, Seasonal estimates of global atmopsheric sea-salt distributions, J. Geophys. Res., 91, 1067-1072, 1986.

Ellis, W. G., Jr., R. Arimoto, D. L. Savoie, J. T. Merrill, R. A. Duce, and J. M. Prospero, Aerosol selenium at Bermuda and Barbados, J. Geophys. Res., 98, 12,673-12,685, 1993.

Ferrare, R. A., R. S. Fraser, and Y. J. Kaufman, Satellite measurements of large-scale air pollution: Measurements of forest fire smoke, J. Geophys. Res. 95, 9911-9925, 1990.

Galloway, J. N., and D. M. Whelpdale, WATOX-86 overview and western North Atlantic Ocean S and N atmospheric budgets, Global Biogeochem. Cycles, 1, 261-281, 1987.

Galloway, J. N., H. Levy, and P. S. Kasibhatia, Year 2020: Consequences of population growth and development on deposition of oxidized nitrogen, Ambio, 23, 120­123, 1994.

Gao, Y., R. Arimoto, M. Y. Zhou, J. T. Merrill, and R. A. Duce, Relationships between the dust concentrations over Eastern Asia and the remote North Pacific, J. Geophys. Res., 97, 9867-9872, 1992.

Gao, Y., R. Arimoto, R. A. Duce, L. Q. Chen, M. Y. Zhou, and D. Y. Gu, Atmospheric non-sea-salt sulfate, nitrate and methanesulfonate over the China Sea, J. Geophys. Res., 101(D7), 12,601-12,611, 1996.

Goudie, A. S., and N. J. Middleton, The changing frequency of dust storms through time, Climatic Change, 20, 197-225, 1992.

Gras, J. L., CN, CCN and particle size in southern ocean air at Cape Grim, Atmos. Res., 35, 233-251,1995.

Griggs, M., Satellite measurements of tropospheric aerosols, Adv. Space Res., 2(5), 109-118, 1983.

Hao, W. M., and M-H. Liu, Spatial and temporal distribution of tropical biomass burning. Global Biogeochemical Cycles, 8(4) 495-503, 1994.

Hameed, S., and J. Dignon, Global emissions of nitrogen and sulfur oxides in fossil fuel combustion 1970-1986. J.Air Waste Manage. Assoc., 42, 159-163, 1992.

Herman, J. R., P. K. Bhartia, O. Torres, C. Hsu, C. Seftor, and E. Celarier, Global distribution of absorbing aerosols from Nimbus-&/TOMS data, J. Geophys. Res., this volume.

Holmes, J., and W. Zoller, The elemental signature of transported Asian dust at Mauna Loa Observatory, Tellus, 48B, 83-92,1996.

Hsu, N. C., J. R. Herman, P. K. Bhartia, C. J. Seftor, O. Torres, A. M. Thompson, J. F. Gleason, T. F. Eck, and B. N. Holben, Detection of biomass burning smoke from TOMS measurements, Geophys. Res. Lett., 23 (7), 745-748, 1996.

Huebert, B. J., T. S. Bates, and N. W. Tindale, Marine aerosol and gas exchange and global atmospheric effects, in Global Atmospheric-Biospheric Chemistry. edited by R. G. Prinn, pp. 39-55, Plenum Press, New York, 1994

Husar, R. B., J. M. Holloway, D. E. Patterson, and W. E. Wilson, Spatial and temporal pattern of Eastern U.S. haziness, A summary, Atmos. Environ., 15, 1919-1928, 1981.

Husar, R. B., and J. D. Husar, Sulphur, in The Earth as Transformed by Human Action, edited by B.L. Tumer et al.. pp. 409-421, Cambridge University Press with Clark University, Cambridge, Mass., 1990.

Husar, R. B., and W. E. Wilson, Haze and sulfur emission trends in the Eastern United States, Env. Sci. Technol., 27, 12-16, 1993.

Iacobellis, S. F., R. Frouin, H. Razaflinpanilo, R. C. J. Somerville, and S. C. Piper, North African savanna fires and atmospheric carbon dioxide, J. Geophys. Res., 99, 8321-8334, 1994.

Ignatov, A. M., L. L. Stowe, S. M. Sakerin, and G. K. Korotaev, Validation of the NOAA/NESDIS satellite aerosol product over the North Atlantic in 1989, J. Geophys. Res., (in press), 1995.

Jennings, S. G., C. D. O'Dowd, T. C. O'Connor, and F. M. McGovern, Physical characteristics of the ambient aerosols at Mace Head, Atmos. Environ., 25A, 557­562, 1991.

Karl, T. R., R. W. Knight, G. Kukla, and J. Gavin, Evidence for radiative effects of anthropogenic sulfate aerosols in the observed climate record, in Aerosol Forcing of Climate, edited by R. Charlson and J. Heintzenberg, pp. 363-382, 1995.

Karyampudi, V. M., and T. N. Carlson, Analysis and numerical simulations of the Saharan air layer and its effects on easterly wave disturbances, J. Atmos. Sci., 45, 3102-3136, 1988.

Kasibhatla, P. S., H. Levy, II, and W. J. Moxim, Global NOx, HNO3, PAN, and NOy distributions from fossil fuel combustion emissions: A model study [NOx HNO3 NOy], J. Geophys. Res., 98 (D4), 7165-7180, 1993.

Kaufman, Y. J., Remote sensing of direct and indirect aerosol forcing, in Aerosol Forcing of Climate, edited by R. Charlson and J. Heintzenberg, 297-332, 1995.

Kiehl, J. T., and B. P. Briegleb, The relative role of sulfate aerosols and greenhouse gases in climate forcing, Science, 260, 311-314, 1993.

King, M. D., Y. J. Kaufman, W. P Menzel, and D. Tanre, Determination of cloud, aerosol and water vapor properties from Moderate Resolution Imaging Spectrometer (MODIS), IEEE Trans. Geosci. and Remote Sensing, 30, 2-27, 1992.

Lacis, A. A., and M. I. Mishchenko, Climate forcing, climate sensitivity, and climate response: A radiative modeling perspective on atmospheric aerosols, in Dahlem Workshop on Aerosol Forcing of Climate, pp. 11-42, May, 1994, Berlin, 1995.

Lee, G., J. T. Merrill, and B. J. Huebert, Variation of free tropospheric total nitrate at Mauna Loa Observatory, Hawaii, J. Geophys. Res. 99, 12,821-12,831, 1994.

Levine J. S., Global Biomass Burning, pp. 47-52, The MIT Press, Cambridge, London,, 1991.

Li, X., H. Maring, D. Savoie, K. Voss, and J. M. Prospero, Dominance of mineral dust in aerosol light scattering in the North Atlantic trade winds, Nature, 380, 416-419, 1996.

Littmann, T., Dust storm frequency in Asia: climatic control and variability, Int. J. of Climatology, 11, 393-412, 1991.

Long, C. S., and L. L. Stowe, Using the NOAA/AVHRR to study stratospheric aerosol optical thickness following the Mt. Pinatubo eruption, Geopys. Res. Letts., 21(20). 2215-2218, 1994.

MacDonald, W. F., Atlas of Climatic Charts of the Oceans, 60 pp., Weather Bureau, Department of Agriculture, Washington, DC, 1938.

Malingreau J. P., G. Stephens, and L. Fellows, Remote sensing of forest fires: Kalimantan and North Borneo in 1982-83, Ambio, 14, 314, 1985.

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 United States, J. Geophys. Res. 99, 1347­1370,1994.

McArdle, N. C., and P. S. Liss, Isotopes and atmospheric sulphur, Atmos. Environ., 29, 2553-2556, 1995.

McClain, P., Global sea surface temperature and cloud clearing for aerosol optical depth estimates, Int. J. Remote Sensing, 10, 763-769, 1989.

Merrill, J. T., Atmospheric long­range transport to the Pacific Ocean, in Chemical Oceanography, edited by J. P. Riley, R. Chester, and R. A. Duce, vol. 10, pp. 15­50. Academic Press, London, 1989.

Middleton, N. J., Dust storms in the Mongolian People's Republic, J. Arid Environ., 20, 287-297, 1991.

Moody, J. L., S.J. Oltmans, H. Levy, II, and J. T. Merrill, Transport climatology of tropospheric ozone: Bermuda, 1988-1991, J. Geophys. Res., 100, 7179-7194, 1995.

Morales, C., ed. Saharan Dust: Mobilization, Transport, Deposition, John Wiley & Sons, Chichester, 1979.

Mukai, H., Y. Ambe, and K. Shibata, Long­term variation of chemical composition of atmospheric aerosol of the Oki Islands in the Sea of Japan, Atmos. Environ., 24A, 1379­1390, 1990.

O'Dowd, C. D., M. H. Smith, and S. G. Jennings, Submicron particle, radon, and soot carbon characteristics over the Northeast Atlantic, J. Geophys. Res., 98 (D1), 1123-1135, 1993.

Olson, 1986.

Ott, S.-T., A. Ott, D. W. Martin, and J. A. Young, Analysis of a trans-Atlantic Saharan dust outbreak based on satellite and GATE data, Mon. Weather Rev., 119, 1832-1850, 1991.

Penner, J. E., Carbonaceous aerosols influencing atmospheric radiation: black and organic carbon, in Aerosol Forcing of Climate, edited by R. Charlson and, J. Heintzenberg, pp. 91-108, 1995.

Penner, J. E., R. J. Charlson, J. M. Hales, N. Laulainen, R. Leifer, T. Novakov, J. Ogren, L. F.Radke, S. E. Schwartz, and L. Travis. Quantifying and minimizing uncertainty of climate forcing by anthropogenic aerosols, Bull. American Meteorol. Soc., 75, 375-400, 1994.

Pham, M., J.-F. M. Muller, G. P. Brasseur, C. Granier, and G. M.Megie, A three-dimensional study of the tropospheric sulfur cycle, J. Geophys. Res., 100 (D12), 26,061-26,092, 1995.

Prodi, F., G. Santachiara, and F. Olioski, Characterization of aerosols in marine environments (Mediterranean, Red Sea, and Indian Ocean), J. Geophys. Res., 88, 10,957-10,968, 1983.

Prospero, J. M., Aeolian transport to the World Ocean, in The Sea, vol. 7, The Oceanic Lithosphere, edited by C. Emiliani, pp. 801-974, Wiley Interscience, New York, 1981.

Prospero, J. M., The atmospheric transport of particles to the ocean, in Particle Flux in the Ocean, edited by V. Ittekkot, P. Schäfer, S. Honjo, and P. J. Depetris, SCOPE Report 57, pp. 19-52, John Wiley & Sons Ltd., Chichester, 1996a.

Prospero, J. M., Saharan dust transport over the North Atlantic Ocean and Mediterranean, in The Impact of African Dust Across the Mediterranean, Oristano, Sardinia, 4-7 October, 1995, Kluwer Academic Publishers, (in press) 1996b.

Prospero, J. M., and T. N. Carlson, Vertical and areal distribution of Saharan dust over the western equatorial North Atlantic Ocean, J. Geophys. Res., 77, 5255-5265, 1972.

Prospero, J. M., and R. T. Nees, Impact of the North African drought and El Nińo on mineral dust in the Barbados trade wind, Nature, 320, 735-738, 1986.

Prospero, J. M., D. L. Savoie, T. N. Carlson and R. T. Nees, Monitoring Saharan aerosol transport by means of atmospheric turbidity measurements, in Saharan Dust: Mobilization, Transport, Deposition, edited by C. Morales, pp. 171-186, John Wiley, Chichester, U.K., 1979.

Prospero, J. M., R. A. Glaccum, and R. T. Nees, Atmospheric transport of soil dust from Africa to South America, Nature, 289, 570-572, 1981.

Prospero, J. M., R. T. Nees, and M. Uematsu, Deposition rate of particulate and dissolved aluminum derived from Saharan dust in precipitation at Miami, Florida, J. Geophys. Res., 92, 14,723­14731, 1987.

Prospero, J. M., M. Uematsu, and D. L. Savoie, Mineral aerosol transport to the Pacific Ocean, in Chemical Oceanography, vol. 10, edited by J. P. Riley, R. Chester and, R.A. Duce, pp. 188-218, Academic Press, San Diego, CA, 1989.

Prospero, J. M., D. L. Savoie, R. Arimoto, and F. Huang, Long-term trends in African mineral dust concentrations over the western North Atlantic: Relationship to North African rainfall, Eos Trans. AGU, 74(43), 146, 1993.

Prospero, J. M., R. Schmitt, E. Cuevas, D. L. Savoie, W. C. Graustein, K. K. Turekian, A. Volz-Thomas, A. D. Diaz, S. J. Oltmans, and H. Levy, II., Temporal variability of summertime ozone and aerosols in the free troposphere over the eastern North Atlantic, Geophys. Res. Lett., 22 (21), 2925-2928, 1995a.

Prospero, J. M., D. L. Savoie, R. Arimoto, H. Olafsson, and H. Hjartarson, Sources of aerosol nitrate and non-sea-salt (nss) sulfate in the Iceland region, The Science of the Total Environment 160/161, 181­191, 1995b.

Pye, K., Aeolian Dust and Dust Deposits, 334 pp., Academic Press, London, 1987.

Pye, K., Processes of fine particle formation, dust source regions, and climatic changes, in Paleoclimatology and Paleometeorology: Modem and Past Patterns of Global Atmosphetic Transport, edited by M. Leinen and M. Sarnthein, pp. 3-30, Kluwer Academic Pubs., Dordrecht, 1989.

Saltzman, E. S., D. L. Savoie, J. M. Prospero, R. G. Zika, and B. Mosher, Elevated atmospheric sulfur levels off the Peruvian coast, J. Geophys. Res., 91, 7913-7918, 1986.

Savoie, D. L., and J. M. Prospero, Particle size distribution of nitrate and sulfate in the marine atmosphere, Geophys. Res. Lett., 9, 1207-1210, 1982.

Savoie, D. L., J. M. Prospero, and R. T. Nees, Nitrate, non-sea-salt sulfate, and mineral aerosol over the northwestern Indian Ocean, J. Geophys. Res., 92, 933-942, 1987.

Savoie, D. L., J. M. Prospero, and E. S. Saltzman, Non-seasalt sulfate and nitrate in trade wind aerosols at Barbados: Evidence for long-range transport, J. Geophys. Res., 94, 5069-5080, 1989a.

Savoie, D. L., J. M. Prospero, and E. S. Saltzman, Nitrate, non-seasalt sulfate and methanesulfonate over the Pacific Ocean, in Chemical Oceanography, vol. 10, edited by J. P. Riley, R. Chester, and R. A. Duce, pp. 219-250, Academic Press, London, 1989b.

Savoie, D. L., J. M. Prospero, S. J. Oltmans, W. C. Graustein, K. K. Turekian, J. T. Merrill, and H. Levy II, Sources of nitrate and ozone in the marine boundary layer of the tropical North Atlantic, J. Geophys. Res., 97, 11,575-1195899 1992.

Savoie, D. L., J. M. Prospero, R. J. Larsen, F. Huang, M. A. Izaguirre, T. Huang, T. H. Snowdon, L. Custals, and C. G. Sanderson, Nitrogen and Sulfur species in Antarctic aerosols at Mawson, Palmer Station, and Marsh (King George Island), J. Atmos. Chem., 17, 95-122, 1993.

Scott, W. D., B. W. Forgan, and J. M. Prospero, Atmospheric turbidity measurements at Broome in western Australia, 1979­1984. J. Royal Society Western Australia, 75, 111­118, 1992.

Shaw, G. E., Aerosols as climate regulators: A climate-biosphere linkage?, Atmos. Environ., 21, 985-986, 1987.

Simkin, T., and L. Siebert. Volcanoes of the World Atlas, Smithsonian Global Volcanism Network, Washington, DC, 1995.

Sirocko, F., and M. Sarnthein, Wind-borne deposits in the Northwestern Indian Ocean: record of Holocene sediments versus modern satellite data, in Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmosphefic Transport, edited by M Leinen and M. Samthein, pp. 401-433 . Kluwer Academic Publ., Boston, 1989.

Stowe L. L., R. Hitzenberger, and A. Deepak, Report on Experts Meeting on Space Observations of Tropospheric Aerosols and Complementary Measurements, WCRP-48, WMO/TD-No. 389,World Meteorological Organization, 1990.

Stowe, L. L., R. M. Cavoy, and P. P. Pelliguino, Monitoring the Mt. Pinatubo aerosol layer with NOAA/II AVHRR data, Geophys. Res. Lett, 19, 159-162, 1992.

Stowe L. L., A. A. Ignatov, and R. R. Singh, Development, validation and potential enhancements to the second generation operational aerosol product at NOAA/NESDIS, J. Geophys, Res., this volume.

Swap, R., M. Garstang, S. Greco, R. Talbot, and P. Kallberg, Saharan dust in the Amazon basin, Tellus 44B, 133-149, 1992.

Talbot, R. W., R. C. Harriss, E. V. Browe, G. L. Gregory, D. I. Sebacher, and S. M. Beck, Distribution and geochemistry of aerosols in the tropical North Atlantic troposphere:relationship to Saharan dust, J. Geophys. Res., 91, 5173-5182, 1986.

Takayama, Y. and T. Takashima, Aerosol optical thickness of yellow sand over the Yellow Sea derived from NOAA satellite data, Atmos. Environ., 20, 631-638, 1986.

Taylor, K. E., and J. E. Penner, Response of the climate system to atmospheric aerosols and greenhouse gases, Nature, 369, 734-737, 1994.

Tegen, I., and I. Fung, Modeling of mineral dust in the atmosphere: Sources, transport, and optical thickness, J. Geophys. Res., 99, 22,897-22,914, 1994.

Tegen, I., and I. Fung, Contribution to the atmospheric mineral aerosol load from land surface modification, J. Geophys. Res., 100, 18,707-18,726, 1995.

Tegen, I., A. A. Lacis, and I. Fung, The influence of mineral aerosols from disturbed soils on the global radiation budget. Nature. 380, 419-422, 1996.

Trijonis, J., R. C. Charlson, R. Husar, W. Malrn, M. Pitchford, and W. White, Visibility: Existing and historical conditions - causes and effects, State-of-Science Technology Report 24 of the National Acid Precipitation Assessment, 1990.

Tsunogai, S., T. Suzuki, T. Kurata, and M. Uematsu, Seasonal and areal variation of continental aerosol in the surface air over the western North Pacific region, J. Oceanogr. Soc. Japan, 41, 427-434, 1985.

Uematsu, M., T. Sugita, V. V. Anikiev, and A. N. Medvedev, Large-scale trasnport of pollution aerosol over the east coast of Asia. Geophys. Res. Letts., 19, 2219-2221, 1992.

Villevalde, Y. V., A. V. Smirnov, N. T. O'Neill, S. P. Smyshlyaev, and V. V. Yakovlev, Measurement of aerosol optical depth in the Pacific Ocean and the North Atlantic, J. Geophys. Res., 99(D10), 20,983-20,988, 1994.

Westphal, D. L., O. B. Toon, and T. N. Carlson, A two-dimensional numerical investigation of the dynamics and microphysics of Saharan dust storms, J. Geophys. Res., 92, 3027-3049, 1987.

Westphal, D. L., O. B. Toon, and T. N. Carlson, A case study of mobilization and transport of Saharan dust, J.Atmos. Sci., 45, 2145-2175, 1988.

Williams, M. A. J., and R. C. Balling, Jr., Interactions of Desertification and Climate, World Meteorological Organization, United Nations Environmental Programme, pp. 270, Arnold, London, 1996.

Wylie, D. P., W. P. Menzel, H. M. Woolf, and K. I. Strabala, Four years of global cirrus cloud statistics using HIRS, J. Climate, 7(12), 1972-1986, 1994.


Figure 1: Radiatively equivalent aerosol optical thickness (EAOT x 1000) over the oceans derived from NOAA AVHRR satellites for the four seasons. The figure incorporates data for the period July 1989 to June 1991.

Figure 2: Aerosol regions over the oceans. The rectangular boxes identify regions for which monthly average EAOT values were calculated (see Table 1 and Fig. 3). Within the rectangles that overly land masses, EAOT values were only obtained over water surfaces.

Figure 3: Monthly mean EAOT for various ocean regions. Note that in 3p, the southern hemisphere annual cycle is shifted by six months so that the seasons can be compared directly for both hemispheres. (Note to reviewers: the symbol for sulfur emissions will be changed in the final version to avoid confusion with the biomass burning symbol.)

Figure 4: Aerosol concentrations in the trade winds at Barbados, July 1989 - June 1991. Daily samples are collected only during on-shore winds. 4a: Mineral dust. 4b: nss-SO4=. Data are plotted against time measured as months from 1 January 1989 [Prospero et al., 1993].

Figure 5: Monthly mean aerosol concentrations in the trade winds at Barbados, July 1989 - June 1991. 5a: Mineral dust; 5b: nss-SO4= [Prospero et al., 1993].

Figure 6: Daily aluminum and nss-SO4= concentrations at Izańa, Tenerife, July 1989 - June 1991. Measurements are made only at night during down-slope wind conditions to ensure that the free troposphere is being sampled. 6a: Aluminum. Concentrations can be converted to an equivalent dust concentration by multiplying by 12.5, assuming that the average concentration of Al in soils is 8%. Data provided by R. Arimoto [see Arimoto et al., 1995]. 6b. nss-SO4= [Unpublished data, D. L. Savoie and J. M. Prospero; see also Prospero et al., 1995a].

Figure 7: Aerosol nss-SO4= concentrations measured on Bermuda, July 1989-June 1991. 7a: Daily concentrations measured during on-shore winds. Data is composited from samples collected at two sites, one on the west end of Bermuda [Arimoto et al., 1995] and a second on the east end; combined, these two stations provide coverage of winds through 335o. 7b: Monthly means composited from the west end and east end sites. To demonstrate the variability over longer time periods, monthly means are shown for two time periods: July 1989 - June 1991 and July 1991 - June 1993. Note the relative stability of concentrations in the fall-winter time periods and the large variability in August. [Unpublished data, D. L. Savoie and J. M. Prospero]

Figure 8: Aerosol nss-SO4= concentrations measured at Mace Head, Ireland, July 1989 - June 1991. 8a: Daily concentrations during on-shore winds. 8b: Monthly mean concentrations for two time periods: July 1989 - June 1991 and July 1991 - June 1993. Note the anomalously high concentrations in December. These are attributed to a few pollution events with unusually high nss-SO4= concentrations; otherwise, concentrations during the winter are relatively low. The December high-pollution events were are not reflected in the AVHRR EAOT distributions in Fig. 1. [Unpublished data, D. L. Savoie and J. M. Prospero].

Figure 9: Aerosol concentrations measured on Midway, 1981 - 1994. Each data point is a one-week-long sample collected during on-shore winds. 9a: Al concentration in aerosols [Prospero et al., 1989; R. Arimoto, personal communication], dust concentrations can be calculated by multiplying by 12.5, assuming an average soil concentration of 8%; 9b: nss-SO SO4= [Savoie et al., 1989b; Savoie and Prospero, unpublished data].

Submit your comments, feedback, questions, and ideas pertaining this page. Your input will be automatically added to the existing annotations. In order to add a new comment, you must be registered with the CAPITA People's Page.