The Asian Pacific Regional Aerosol Characterization Experiment (ACE-Asia) intensive phase took place from 30 March to 4 May 2001, involving extensive aerosol observations from aircraft, ship, and ground stations. Meteorological events of significance to the ACE-Asia intensive observations made during this period were identified and analyzed using a combination of gridded meteorological analysis fields, observations, and analyses of precipitation, atmospheric radiation distributions, and other variables, and products prepared from these data. The observed flow characteristics were found to be close to the center of the range of interannual variation, with most anomalies spatially confined and small in magnitude. The occurrence and distribution of midlatitude cyclones which had an impact on the circulation and transport patterns affecting the area of the campaign were close to the climatological mean. Total aerosol scattering and the fraction caused by submicron particles, and trajectory analysis results from points along the flight track of the C-130, were used to examine the relationship between the aerosol distributions and transport patterns and the meteorological situation. Notably strong coarse-mode scattering was observed at low altitudes in post cold-frontal circulations during dust outbreaks, and very strong mixed-mode scattering occurred in plumes from polluted areas flowing offshore removed from important surface or upper air features.
 In this paper we discuss some of the meteorological aspects of the Asian Pacific Regional Aerosol Characterization Experiment (ACE-Asia) intensive observation period, which was carried out from 30 March to 4 May 2001. The motivation for the campaign, its relationship to earlier International Global Atmospheric Chemistry programs, the equipment used and the strategies employed were discussed by Huebert et al. . Briefly, the goals were to quantify the spatial and vertical distributions of aerosol concentrations and properties and to determine the column integrated radiative effects of these aerosols.
 A number of authors have treated various aspects of the problems addressed here. Merrill et al.  presented an overview of the meteorological conditions during the Pacific Exploratory Mission—West, phase B campaign, which took place in winter, 7 February to 15 March 1994. Fuelberg et al.  give a meteorological overview of the Transport and Chemical Evolution Over the Pacific intensive period, 23 February to 9 April 2001; that multiaircraft campaign included some aerosol observations but emphasized gas-phase photochemistry. A model-based assessment of the transport of pollutants out of Asia has been presented by Liu et al. .
 The primary results and the event-specific figures in this paper are based on the operational version of the NOGAPS model [Hogan and Rosemond, 1991; Hogan and Brody, 1993]. This is a global spectral model in hybrid-σ vertical coordinates. The fields available for this project were at 1° resolution at 13 levels between 1000 hPa and 150 hPa, with enhanced resolution in the lowest 3 km. The fields in the archive at the isobaric levels include the geopotential height, wind components, temperature, dew point depression, grid-scale pressure tendency (vertical velocity) and total and convective cloud amount estimates. The model is run in a continuous data assimilation/forecast cycle, and the fields used here are analyses at 0000 and 1200 UTC, and 6-hour forecasts at 0600 and 1800 UTC each day. Standard dynamical meteorological principles were used to relate the mass distribution of dry air, i.e., the geopotential, the water vapor mixing ratio (determined from the dew point depression) and the wind vector to derived quantities such as convergence and divergence.
 The NCEP/NCAR Reanalysis project fields [Kalnay et al., 2001; Kistler et al., 1996] were used in monthly mean calculations and figures, and in event-specific form for intercomparisons with the other meteorological data. The resolution is lower in both the horizontal and vertical than the NOGAPS fields: 2.5° × 2.5° and eleven levels up to 150 hPa of seventeen levels overall, with 4 levels up to 700 hPa. The fields are available as analyses at 0000, 0600, 1200 and 1800 UTC daily. Mean and anomaly fields for column integrated quantities such as precipitation and outgoing long wave radiation were obtained from the Web site of the NOAA-CIRES Climate Diagnostic Center, available at http://www.cdc.noaa.gov. Also, additional information was used from the Climate Prediction Center Climate Diagnostics Bulletin; the Center is a branch of the NOAA National Centers for Environmental Prediction, and the Bulletin is available on the Web at http://www.cpc.ncep.noaa.gov.
 We also made use of meteorological and other fields from the RIAM-CFORS mesoscale forecast and chemical transport model [Carmichael et al., 2003; Uno et al., 2003]. Back trajectories, calculated kinematically using the three-dimensional winds, were prepared for points along the flight track of the C-130 and Twin Otter aircraft, at intervals of 3 min during the research flights, and along the track of the R/V Ron Brown and the surface site at Gosan, Jeju Island, Korea. The trajectories extended 5 days back in time, or until the hypothetical parcel left the model domain.
2. Meteorological Setting
 Both the mean and transient circulations contribute to patterns of aerosol transport, as the structure of the mean flow determines the overall velocity distribution, while this and perturbations cause the occurrence of migratory meteorological systems which in turn cause plumes and voids to cross the mean stream lines. Here we examine in turn the time-average circulation, the representativeness of the flow encountered during the field program, the observed precipitation and the occurrence of transient midlatitude storms and high-pressure centers.
2.1. Mean Horizontal Winds
 The mean streamlines and isotachs are shown at 500, 700, and 1000 hPa in Figure 1. The slight cyclonic curvature apparent in the 500 hPa streamlines is the net effect of the succession of cyclonic and anticyclonic disturbances passing through this area, with the former somewhat stronger on average than the latter. The mean wind speeds at 500 hPa exceed 20 ms−1 equatorward of 30°N and in the area south and east of Japan, where the meridional temperature gradient is most consistently strong. The 700 hPa and 1000 hPa isotachs exhibit areas of higher mean wind speed around 45°N, also an effect of transient cyclones.
 A height versus latitude cross section of the zonal wind, averaged over the month of April 2001 at 120°E, is shown in Figure 2. The main features are the midlatitude westerly tropospheric wind maximum near 30°N and the strong easterly winds in the upper troposphere in the tropics.
 The jet stream which dominates Figure 2 exhibits slight variations in structure and intensity between 105° and 135°E (not shown). The strongest mean winds occur just west of the continental boundary, at 110°–125°E and reach 50 ms−1; the mean winds at 105°E are slightly weaker because the height of the maximum varies more at this location. The core of the jet as observed near 200 hPa, ≃12 km, is close to the climatological mean maximum wind height. Stronger maximum winds, but otherwise similar characteristics were observed earlier in 2001 during TRACE-P. The mean maximum jet stream winds were about 65 ms−1 for the first half of that campaign, 23 February to 17 March, peaking farther offshore than during the ACE-Asia period, over 120°–135°E [Fuelberg et al., 2003]. Similarly, the maximum speed in the jet, averaged over the PEM-West B period, 7 February to 15 March 1994, was ≃70 ms−1 [Merrill et al., 1997]. The core wind speeds decrease from late winter into spring because of the seasonal decrease of the meridional temperature gradient.
 Tropical circulations had undergone a relatively rapid transition from cold-phase (La Niña) to near-normal conditions. The tropical sea surface temperatures were near normal, and the Southern Oscillation Index at −0.1 was very close to normal, although positive anomalies were present in convection and in Outgoing Long-wave Radiation in the western tropical Pacific. Extratropical SST anomalies were present, but are not a dominant influence; positive anomalies ranging as high as 2.5°C and averaging approximately 0.5°C were present in the area 30–40°N, 125–150°E. Anomalies in the OLR over Japan and the Sea of Japan were consistently positive but insignificant, less than 10 W m−1. In the extra-tropics the observed flow characteristics were also close to the climatological mean conditions. The 200 hPa stream function was near its long-term mean, here assessed using the 1979–1995 average. Slight anticyclonic anomalies were present near the area of ACE-Asia observations, but these were not significant. The 500 hPa heights were within normal ranges over the ACE-Asia area, with slight positive anomalies of 30–40 m in the 564–546 dm range (approximately 30–50°N) in an elongated area from 105°E to 125°E. Increased heights at this level reduce the incursion of progressive troughs and thereby weaken or forestall development of midlatitude storms. More significant positive 500 hPa height anomalies were present elsewhere in the global subtropics. The anomalies in the ACE-Asia area were minimal and, as noted further below, the occurrence of storms was very close to climatological mean values.
 In Figure 3 the observed rainfall for April 2001 and the deviation from an estimated long-term mean distribution are shown, as estimated by the Global Precipitation Climatology Project [Adler et al., 2003]. The band of maximum precipitation near 30°N and east of 125°E is the result of midlatitude storms strengthening as they pass over the warm waters of the Kuroshio current there. Midlatitude storms are also the source of the less intense area of precipitation north and east of Japan. Both midlatitude storms and subtropical convective disturbances contribute to the secondary maximum in precipitation in southeast China from 22°N to 33°N. The GPCP composites combine microwave and infared satellite estimates with gauge data to form a 2.5° × 2.5° gridded field. Precipitation distributions estimated with other approaches are consistent overall but differ in detail from those shown here. The mean and anomaly estimates fall within the range of interannual variation assessed over the previous ten year period in the GPCP data set, and the differences between these estimates is believed to be inconsequential. The maximum amounts, southeast of Japan in the range 25–30°N, and the positive anomaly values relative to the 1979–2000 base period of 2–6 mm d−1 there may be significant, but their impact on the aerosol distributions as measured during ACE-Asia is likely to be limited because transport pathways which were sampled crossed this area infrequently. The deviation in the band of below normal precipitation in 35–45°N east of 120°E, although smaller in magnitude, is more likely to have impacted both the observed aerosol distributions and the transport to the open Pacific Ocean and beyond, because it covers part of the sampling area and also covers important transport pathways out of Asia.
 Another indication of the representativeness of the meteorological conditions is the occurrence of transient disturbances. The tracks of surface cyclones and anticyclones affecting the ACE-Asia area are shown in Figure 4. We include anticyclones in this analysis specifically because of the emphasis on measuring profiles of radiative characteristics in this campaign; the reduced occurrence of clouds in anticyclonic flows resulted in more frequent aircraft sampling in anticyclonic than cyclonic circulations. Cyclones form downstream (east) of vorticity maxima in the upper-level flow and, guided by midtropospheric winds tend to move toward the east initially and toward the northeast when east of 135°E. This pattern can be seen in the top panel of Figure 4. Midlatitude storms often strengthen or regenerate when they pass over the warm waters of the Kuroshio current southeast of Japan; the cyclones south of 35°N in the top panel developed in this way. The frequency and distribution of cyclones encountered during the campaign is consistent with the long-term average occurrence of such systems. The spatial and seasonal distribution of the initiation and occurrence of cyclones was studied by Chen et al. ; rapidly deepening systems offshore have also been examined [Chen et al., 1992]. The relative occurrence of anticyclonic conditions cannot be accurately assessed by inspection of Figure 4 because only migratory centers exhibiting closed circulations over areas smaller than approximately 10° × 10° were tracked in this analysis. Broader areas of high pressure were also commonly present, especially over the interior continental area and also on the equatorward side of the polar and subtropical frontal boundaries in the typical midlatitude cyclone family conformation. Steady features were not tracked here, and also not captured by this analysis are short-lived features identifiable for 36 hours or less, whether slow moving or obviously migratory, and features which are broad or amorphous. Another perspective on transient features and episodes during the campaign is presented in the next section.
2.3. Meteorological Events and Relationships to Aircraft Sampling
 The latitude versus time sections in Figure 5 document the passage of the most prominent transient features into the area in which much of the sampling took place, based on juxtaposed NOGAPS analysis-forecast fields. Progressive troughs and ridges at 500 hPa are manifest as groups of contours open toward the pole and toward the equator, respectively, and circulations which result in a height maximum or minimum cut off from the downstream flow appear as closed contours. The 1016 and 1000 hPa contours outline areas of notably high and low sea level pressure, respectively. Also, hachuring is used along the 1008 hPa line, against the pressure gradient. Corresponding sections made from the Reanalysis fields (not shown) exhibit similar patterns, taking account of the lower spatial resolution of that dataset and the impact of the analysis-forecast cycle on the NOGAPS fields illustrated here.
 A migratory anticyclone passed 120°E over 30 March to 1 April, supported by the upper level ridge evident in the 500 hPa section on Figure 5. The closed 1028 hPa isobar crossed 120°E near marker 1. Sampling on C-130 Research Flight 1 was partly on the southeast side of this broad high. The first of a series of significant dust outbreaks had begun at 1800 UTC on 5 April near the Mongolian border. Widespread airborne dust accompanied the low-pressure center which crossed 120°E at 988 hPa near 50°N at 1200 UTC on 7 April; this is marked 2 on Figure 5. The upper-level system had become quasi-barotropic by the time it crossed the continental boundary; the corresponding 522 dm cutoff low at 500 hPa crossed 120°E at 0600 on 7 April. The consequent reduction in dynamic forcing allowed the storm to weaken, and when it passed 135°E at 1800 UTC on 8 April the surface disturbance had filled to 1009 hPa. The upper-level ridge which passed 120°E north of 54°N at 0000 UTC on 9 April accelerated with the advance of a 534 dm cutoff low at 0000 UTC on 10 April, near 50°N. The second surface system of the sustained dust outbreak strengthened farther east, but the 1008 hPa isobar crossed 120°E on 10 April, marked 3 on Figure 5. This sea level system crossed 135°E at 1800 UTC on 10 April at 997 hPa, and the frontal passage associated with this system at Jeju took place at 1300 UTC on 11 April. Sampling in the Yellow Sea on 11 April, on C-130 Research Flight 6, encountered dust from this complex, together with fresh pollution aerosol from Qingdao, China. Mesoscale meteorological and dust transport modeling provide a more detailed picture of the structure and other characteristics of this system [Liu et al., 2003].
 The 1023 hPa anticyclone which passed 120°E near 26°N at 0600 UTC on 12 April, labeled 4 on Figure 5 moved slowly offshore and weakened over the South China Sea. Aerosol scattering profile characteristics from C-130 Research Flight 7 in the associated postfrontal circulation are discussed in section 3.3, below. On 13 April one of several multiplatform intensive observation efforts was completed in nearly undisturbed conditions near Oki Island in the Sea of Japan [Kahn et al., 2004]. The low center marked 5 on Figure 5 formed at 120°E and deepened slightly as it moved eastward, crossing 135°E at 1001 hPa at 1200 UTC on 23 April. The anticyclone behind the associated cold front as it advanced over the Yellow Sea is marked 6 on Figure 5, and the results of sampling in this area on 24 April are discussed below in section 3.4. A weak low-pressure center formed east of 120°E on 30 April, near marker 7, supported by the migratory trough evident in the 500 hPa height cross section on Figure 5. C-130 Research Flight 16 circumnavigated the sea-level system, as discussed below in section 3.5. While these events dominated variations in the circulation, the observed distribution of scattering and the underlying transport of aerosols involve additional factors, as discussed in the next section.
3. Aerosol Scattering Profiles Relating to Transient Features
 Here we illustrate and discuss the range of aerosol scattering observed as this relates to the transient circulation, specifically in the several transport domains around the midlatitude cyclones and anticyclones. We make use of total scattering at 550 nm, σsp, estimated from integrating nephelometer data covering the angular range 7–170°. This is expressed as an extinction coefficient, the product of a scattering efficiency, m2g−1 and a concentration, μg m−3, yielding (106 m)−1 or M m−1. Thus σsp quantifies the amount of light removed per million meters of viewing distance. We also consider the fine mode fraction, FFscat, the ratio of the scattering from the total aerosol stream to that from the submicron particles, again at 550 nm [Anderson et al., 2003]. Their analysis yielded campaign-wide average values for σsp of approximately 83 M m−1, with approximately 31 M m−1 arising from submicron particles. We designate scattering as fine-mode dominated when FFscat > 0.6, the criterion used by Anderson et al. . Also included in the analysis is the potential temperature of the air sampled, used as an indicator of the distribution of static stability.
3.1. Transport Patterns in the Warm Sector of Midlatitude Storms
 A 997 hPa low-pressure center crossed 135°E near 45°N at 0600 on 2 April, supported by the cutoff low at 500 hPa which had passed 120°E north of 50°N the previous day (Figure 5). Sampling in the Sea of Japan area on Research Flight 2 of the C-130 was in the warm sector of this system. Trajectories (not shown) approach from the southwest in the lower layers, with cyclonic curvature. The boundary layer flow was most recently over the continent north of Shanghai, and the lower to midtropospheric flow was more nearly westerly, and ascending. Aerosol scattering profiles from Research Flight 2 are shown in Figure 6. In the well-mixed layer extending above 1 km there is strong scattering, σsp > 100 Mm−1 from fine mode particles. The weaker scattering above 2 km is consistent with the trajectories remaining well above the surface as the air passed over eastern China, and with intermittent precipitation west of 110°E where the air passed two days earlier.
 At 0600 on 18 April a 997 hPa low formed just west of 135°E at 47°N. The central pressure remained unchanged as this system moved past 135°E at 1200 UTC. It stalled there until an upper-level trough approached at 0600 on 19 April, after which it strengthened and moved toward the northeast. Sampling early in C-130 Research Flight 10 encountered the cyclonic flow around this system before it had fully formed. The aerosol scattering profile in Figure 7 indicates significant scattering by coarse-mode particles below ∼1500 m, decreasing through a stable layer to weak scattering by mixed and fine-mode particles above 2 km. Back trajectories (not shown) indicate flow crossing the Shanghai area sampled in the middle boundary layer, contrasting with a near-surface flow from the south. In the troposphere above the boundary layer the flow is convoluted but generally from the southwest. Above 2500 m altitude the transport is approximately westerly.
3.2. Transport Patterns Near a Weak Anticyclone
 Profiles from C-130 Research Flight 5 are shown in Figure 8. There was significant scattering in the stable boundary layer below 500 m, and variable scattering in the weakly stratified layer extending up to ∼2 km, all dominated by fine mode scattering. Above this a weakly stratified layer approximately 600 m thick exhibited weak total scattering in the lower 300 m and strong scattering by increasingly coarse mode particles in the upper 300 m. A weak surface anticyclone was in place over the Sea of Japan, with a broad ridge with its axis near 140°E at 500 hPa. The trajectories for this profile in Figure 9 show that strong vertical shear in the transport direction and speed suggest sources in nearby Japan for the lowest layer and the possibility that air which had been primarily over the ocean for more than a few days was sampled in the 0.5–1 km range.
3.3. Transport Patterns in the Postfrontal Environment
 Aerosol scattering profiles for C-130 Research Flight 7 are shown in Figure 10, in the Yellow Sea area, in the dry air behind the cold front associated with the first low-pressure center of the major dust outbreak. Very strong scattering below 500 m was due to mixed mode aerosol in a stratified layer, above which the scattering fell off gradually to σsp < 50 M m−1 above 2500 m. Above this there was weak scattering dominated by the fine mode, and Redemann et al.  reported Sun photometer derived aerosol optical depth estimates yielding a similar analysis of this distribution. Trajectories for this profile are shown in Figure 11. The transport below 500 m was from the west and northwest and at low altitudes. The flow from the northwest above 2500 m altitude originated in the dust-free area well northwest of the widespread dust outbreak, accounting for the relative absence of coarse-mode scattering at these altitudes.
 Some sampling in postfrontal conditions also took place during C-130 Research Flight 11, when the C-130 was in the Sea of Japan, near 40°N, 140°E. There was adequate dynamical support in the upper tropospheric circulation, but the surface disturbance had not fully developed and the frontal circulation was weak in this case. Transport from the northwest brought mixed-mode aerosol with total scattering values much smaller than those observed on Research Flight 7, σsp ∼ 10 M m−1 in a mixed layer over 1500 m deep.
3.4. Transport Patterns in Relation to an Advancing Cold Front
 In Figure 12, height versus latitude cross sections of the meteorological parameters along 124.5°E are shown for 0000 UTC, 24 April 2001. The features of interest occur at mark 6 on Figure 5 as discussed above in section 2.3. Selected trajectories for the 24 April C-130 Research Flight 13 are shown in Figure 13, and here we discuss the relationship between the observed aerosol characteristics, these trajectories, and the features in the cross section. In Figure 12 a weak cold frontal band near 25°N is indicated by the closely packed isentropes (top left panel), the enhanced ageostrophic winds (bottom left panel), and the more northerly surface wind north of 25°N (bottom right panel). A broad band of clouds accompanied this front, which was associated with a surface low-pressure center located near 31°N, 130°E which developed slowly in place before moving off to the east after 0600 UTC on 25 April. There is a shallower frontal band near 39°N, indicated on the cross section by the wind shift, ageostrophic winds and complex stability structure. The flight legs along 124.5°N took place between these features. The midlevel and upper-level trajectories shown in Figure 13 approach the aircraft track from the west and northwest, on fluctuating winds associated with a transient lower-tropospheric trough. The contrast in two of their paths back to the continent arose from a migratory anticyclone which had passed offshore from Shanghai two days earlier. Along the cross section the surface-level winds are from the northeast north of 32°N and weakly from the northwest north of 40°N; these winds were on the southeast flank of an anticyclone over the continent, a 1029 hPa high centered near 110°E (not shown). The trajectories north of 35°N show weakly subsiding, anticyclonic flow associated with this center. Very heavy dust loading was encountered during the low-level legs north of 37°N, with coarse-mode dominated total aerosol scattering in excess of 650 M m−1 over a widespread area. C-130 Research Flight 14, which took place on the 25th, also encountered this feature. The low-level flow pattern was dominated by a very broad area of high pressure, which extended from 28°–45°N, 107°–125°E at 0000 UTC on 25 April. The aircraft encountered mixtures of dust and pollution on low and midlevel legs in this area, with total aerosol scattering varying up to 150 M m−1. In the trajectories for flight legs near 3600 m altitude the influence of a transient trough west of 120°E the previous day can be seen, while at altitudes above 4500 m the flow was nearly zonal. Similar total aerosol scattering values were observed above 1.5 km, up to the highest altitudes sampled in this area, ∼5500 m.
3.5. Transport Patterns in Relation to a Midlatitude Storm
 In Figure 14, trajectories are shown at selected points along the flight track of the C-130 during Research Flight 16 on 30 April 2001. Clouds and precipitation were present along a frontal boundary in the area encompassed by the flight track, associated with a weak low-pressure center near 32°N, 135°E, noted as label 7 in Figure 5, above. The cyclonic flow around the low was quite shallow, with closed circulations limited to heights below 2.5 km during the period of the aircraft sampling. The dynamical forcing of this storm was limited prior to 1800 UTC on 30 April, and its initial development was quite slow: the central pressure decreased from 1006 to 1001 hPa from 0000 to 1800 UTC as it moved to 141°E.
 In the top left panel of Figure 14, trajectories for the first 2 hours of the flight are shown. The 1519 m altitude segment was part of an intercomparsion leg with the ARA King Air aircraft. The total scattering was low in this area, σsp ∼ 30 M m−1. Farther along the southward flight track the low-level flow was fully anticyclonic, as indicated by the trajectories for altitudes lower than 200 m. Similar paths are indicated by the two low-level trajectories in the top right panel of Figure 14, well within the warm-air sector of the storm. Sampling in that area indicated very clean conditions in both the aerosol and gas phase: low scattering and absorption in both submicron and coarse mode particles, σsp < 10 M m−1, O3 < 10 ppbv and CO < 100 ppbv. The trajectory at 1701 m shows the influence of a transient anticyclone which had passed through that area three days before the flight, a subsiding, clockwise flow. The trajectories near 3 and 4 km altitude show weaker anticyclonic flow. Clean conditions were also encountered in low-level flight legs shown in the bottom left panel of Figure 14. North of 25°N, however, the transport path illustrated by the trajectory at 1078 m was associated with significant scattering in layers, σsp ∼ 40 M m−1. Taking uncertainties in to account the material could have come from the mainland or from Chinese Taipei. Farther north the aircraft was flown below cloud base and encountered the strongest submicron aerosol scatting of the campaign, among numerous other pollution indicators. The aerosol was of mixed mode and the total scattering exceeded 400 M m−1, with the submicron total aerosol scattering exceeding 250 M m−1. The trajectory at 219 m confirms that this was the plume from the Shanghai region. The concentrations fell off farther north, but SO2, CO and CO2 had peak values in the cloud layer north of 28°N, with the transport path illustrated by the trajectory at 537 m in the bottom right panel of Figure 14. Lower amounts of material from Shanghai were found in the area northeast of 29°N, 125°E, along transport paths illustrated by the trajectory at 241 m. The reduced aerosol loading could be because of reduced pollution sources in the area north of 34°N, because of the longer travel time from the continent to the sampling area, or a combination of these. The trajectories at altitudes above 1500 m arrive at the flight track from the northwest, but are in the strong westerly flow, and originate south of the sampling location and well west of 115°E.
 The concept of airstreams which resulted from isentropic trajectory analyses led to improved understanding of the relationship between midlatitude disturbances and cloud masses [Carlson, 1980] and precipitation distributions [Browning, 1986]. More recently the application of the concept of conveyer belts introduced by these authors has proved to be informative in the analysis of aeolian transport patterns and the resulting distributions of photchemically active compounds, particularly over the oceans where the absence of fresh emission of pollutants helps the imprint of upstream continental sources stay intact [e.g., Cooper et al., 2002; Liu et al., 2003]. There are two factors which tend to lessen the utility of the conveyer belt picture in ACE-Asia. First is the fact that so much of the aircraft sampling occurred far from the cyclone centers, away from the area where the conveyer belts are most active. Second is the combined influence of condensed water and precipitation on the transformation and removal of aerosols, initiated by the upward vertical motion in both the warm and cold conveyer belts. Of course heterogeneous processes also have a role in gas-phase chemistry of ozone precursors, but for aerosols the largest direct radiative effects and export fluxes will not be found in the vicinity of the ascending airstreams.
 As detailed above in section 3, the aircraft sampling occurred more frequently in anticyclonic than in cyclonic flow conditions. Five of the 19 C-130 research flights were at least partially impacted by the circulation around midlatitude storms; two of each of these were aerosol and radiation-oriented flights, and one flight had a combined objective. The remaining 14 flights were conducted in anticyclonic conditions: two flights with combined objectives, five with primary goals related to radiation and seven with aerosol survey or characterization goals. The vertical distribution of the aerosols observed on these flights was influenced by this selective sampling of the environment. Dust and other coarse-mode aerosols must undergo a lifting process in order to be transported significant distances from the source area. The subsidence which accompanies anticyclonic flow, while slow, gradually decreases the mean altitude of layers of both fine-mode and coarse-mode aerosols as time passes. This is in contrast with the effects of synoptic-scale uplift and the more frequent cloud-mediated upward transport in areas of cyclonic circulation. The vast majority of fine-mode pollution was observed in the boundary layer in ACE-Asia, while significant dust concentrations were found throughout the lower troposphere [Anderson et al., 2003]. Sampling preferentially in areas of downward vertical motion, away from the uplifting effects of clouds and fronts strengthened, or may have controlled this contrast.
 Gridded meteorological fields from global analyses, trajectory analyses based on wind fields from a mesoscale model, and remotely sensed estimates of rainfall and other parameters were used to describe the occurrence, motion and fate of disturbances for the ACE-Asia operations area during the April 2001 campaign. Further, the data were compared with corresponding values from climatological archives, indicating that the circulation and other conditions were close to long-term means. Case studies of flow influenced by midlatitude storms and subsynoptic scale features which were sampled comprehensively by the C-130 aircraft were described. Notably strong coarse-mode scattering was observed at low altitudes in post cold-frontal circulations during dust outbreaks, and very strong mixed-mode scattering occurred in plumes from polluted areas flowing offshore removed from important surface or upper air features. The observation that fine-mode aerosol scattering was observed predominantly in the lower layers of the atmosphere was related to the selective sampling distribution, with more frequent observations in anticyclonic than cyclonic conditions.
 The work at the University of Rhode Island was supported by NSF grant ATM-0002227. Computing facilities and data archives of the National Center for Atmospheric Research were used. Data access assistance was provided by the Joint Office of Science Support, UCAR. NCAR and JOSS are also supported by NSF. We acknowledge George Huffman, David Bolvin, Robert Adler, and the GPCP project for provision of the GPCP combined precipitation data. We acknowledge with gratitude the assistance of G. Kurata of the Department of Electrical Engineering, Toyohashi University of Technology, Japan, who provided trajectory analysis results based on meteorological fields from the RIAM-CFORS regional numerical model simulations. This work benefited from ACE-Asia investigators who shared their data and scientific perspectives. Byron Blomquist, Tad Anderson, and Sarah Doherty are especially acknowledged. The authors are also grateful for the constructive comments provided during the review process.