| 
| 
| 
| 
|
| 
| 
| 
| 
|
| 
| 
| 
| 
|
| 
| 
| 
| 
|
| 
| 
| 
| 
|
Manuscript to be submitted to the JGR
PowerPoint presentation at: http://capita.wustl.edu/Central-America/reports/JGR/SmokeImpactJGR.htm
April, 2000
The Impact of the 1998 Central American Smoke on the Atmospheric Environment of Eastern North America
R. B. Husar[1], B. A. Schichtel1, S. R. Falke1, F. Li1, W. E. Wilson[2], J. Pinto2, W. C. Malm[3], D. G. Fox[4], G. C. Feldman[5], C. McClain5, N. Kuring5, B. N. Holben5, E. F. Vermote[6], J. R. Herman5, C. D. Elvidge[7]
Abstract. During a ten-day period, May 7-17, 1998, smoke from numerous widespread fires in Central America drifted northward and caused severe perturbation of the atmospheric environment over parts of Eastern North America. On May 14 and 15 the 4000 km contiguous smoke plume stretched from Guatemala to Hudson Bay in Canada. The location of fires was detected by the DMSP and AVHRR satellite sensors, the smoke's spatial and temporal distribution was derived from the TOMS satellite and surface visibility measurements, and the aerosol spectral reflectance was derived from the SeaWiFS satellite data. The detailed aerosol properties were determined from surface based measurements of size segregated mass concentration and chemical composition, horizontal extinction coefficient and vertical spectral optical thickness. Along the smoke plume the visual range was below 5 km, and aerosol concentrations were above 100 mg/m3 (> 500 umg/m3 in Brownsville, TX). At least 70% of the smoke aerosol was below 2.5 mm in size and composed of organics (50%), elemental carbon (2%) and an inexplicably high (26%) fraction of sulfate. The smoke extinction efficiency was about 4 m2/g and the characteristic size inferred from spectral optical thickness was about 0.6 mm, which produces the ‘blue moon’ phenomenon. The smoke aerosol has reduced the amount of solar radiation reaching the surface and increased the surface albedo by 10-30% in the near UV. The radiative perturbation in the near IR was minimal. An unexplained feature of the smoke is the yellowish coloration since it is not known whether it is due to pure scattering or preferential absorption in the near UV.
Throughout the spring of 1998, thousands of fires burned throughout Central America. Springtime biomass burning is common in that region, but the 1998 fires were much more intense than in other years. Thick smoke plumes lingered over southern Mexico, Guatemala, Honduras, and the adjacent Gulf of Mexico and Eastern Pacific. In several multi-day events, there were also major smoke incursions from Central American into Eastern North America with dramatic impact on the atmospheric environment over the region. This is a summary report of the 1998 Central American smoke event and its impact on the particulate matter (PM) and ozone air quality as well as on the atmospheric radiation pattern over Eastern North America. This paper does not specifically address the transport of the smoke plumes. This aspect of the event has been covered by other organizations that have adequately modeled the plume's transport [ARL, 1998]
The atmospheric effects of biomass burning has received increasing interest from the research community with focus on savanna fires in Africa [Andreae et al., 1996], deforestation fires in South America [Artaxo et al., 1998], and forest fires in Indonesia [Radojevic and Hassan, 1999]. Biomass burning over Central America is also a well-known source of smoke in the region. In fact, one of the first published reports on satellite remote sensing of aerosols was an illustration of smoke plumes emanating from Central America to the Pacific [Parmenter, 1971]. More recent aerosol remote sensing studies have revealed elevated aerosol thickness (AOT) over the oceans surrounding Central America [Herman et al., 1997; Husar et al., 1997].
Unlike in the earlier years, the atmospheric research and air quality management communities followed the 1998 Central American fires with keen interest using a variety of UV, visible and infrared satellite remote sensors. The atmospheric science community was primarily interested in the effects of smoke on the radiative budget [Christopher et al., 2000] and other effects on the atmospheric environment such as enhanced lightning in smoke plumes [Lyons et al., 1998]. In addition, smoke plumes can be transported over thousands of kilometers, and they aid the understanding of transport, transformation and removal processes along their path through visualization. Hence, major fire events provide unique opportunities to examine the inner workings of the atmospheric system and help establish the link between local and global scale atmospheric processes.
Major forest fires are often associated with catastrophic
consequences to humans and their environment. During the 1998 fires, the
interest of the air quality management community was in protecting the public
health through health advisory bulletins.
Knowledge about the fires and smoke is required in at least three
phases, or aspects, of air quality management: (1) People need to be warned and
evacuated from the danger zone in the vicinity of active fires (2) Downwind of
the fire zone (up to 1000 km), the smoke aerosol concentrations may be
sufficiently high that ‘health advisories’ may need to be issued; (3) The smoke
may elevate the ambient aerosol concentration to near or above the level of
daily maximum ambient standards, 140 mg/m3 for PM10 and 65 mg/m3 for PM2.5. In
many instances the origin of smoke is outside the jurisdiction of state and
federal air pollution control agencies.
The Central American fires were a clear case for such extra
jurisdictional impact. In fact, one of the purposes of this study was to support
EPA and the U.S. states in deciding which states should get exceedance
waivers due to the Central American smoke event.
The unpredictability of these fire events present a unique challenge to episode-oriented atmospheric research. The study had to rely largely on routine monitoring data and some serendipitous observations. In fact, this report is based exclusively on routine environmental monitoring data that were collected by many agencies and shared with the research community through the Internet.
Satellite remote sensing was crucial in detecting the fire locations, identifying the smoke plumes, following their evolution, and characterizing the spatial, temporal, and optical characteristics of the moving smoke aerosol. Surface-based monitoring of various aerosol properties and ozone concentrations provided detailed smoke properties at a few select locations.
During the 1998 fire season, the fire locations were reported daily using the nighttime visible images from the Operational Linescan System (OLS) sensor, on the Defense Meteorological Satellite Program DMSP (http://www.ngdc.noaa.gov/dmsp) meteorological satellite. Fires were distinguished from the nighttime city lights by their transient behavior [Elvidge et al., 1996]. The fire locations were also identified using multi-spectral data from the NOAA Advanced Very High Resolution Radiometer (AVHRR) (http://www.osei.noaa.gov) sensor. The AVHRR-derived fire product uses the brightness at 4 mm developed by Kaufman et al., [1998]. The current fire detection methods indicate the locations of fires but they cannot quantify the intensity of the fires and the associated smoke emissions.
Operational retrievals of absorbing aerosol index from the Total Ozone Mapping Spectrometer (TOMS) satellite [Herman et al., 1997] (Figure 1) provided useful information on the daily spatial distribution of the smoke for the post-analysis. The TOMS absorbing aerosol signal is a semi-quantitative index of the columnar absorption by aerosols at 0.34 mm. The signal is derived from the absorption of the upwelling Rayleigh scattering in the lower strata of the atmosphere. The columnar absorbing aerosol index (AAI) was used to locate the smoke sources and to estimate semi-quantitatively the magnitude and distribution of smoke aerosol plumes. The daily gridded AAI data were obtained from the NASA TOMS (http://jwocky.gsfc.nasa.gov/aerosols/aerosols.html) project website.
Spectral reflectance data from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) (http://seawifs.gsfc.nasa.gov/SEAWIFS.html) sensor [McClain et al., 1998; Barnes et al., 1999] provided detailed spatial pattern of the smoke at local noon each day. On the SeaWiFS images the smoke cloud was recognized by its yellow coloration near the source, partial transparency and smooth spatial texture compared to white-gray and highly textured clouds (Figure 2a-d). The color images of SeaWiFS provided a rich visible context for the radiative effects of the smoke, including the magnitude of spectral backscattering, relationship to clouds and the spectral albedo of the underlying landmasses.
The raw (Level 1A) Local Area Coverage (LAC), 1 km resolution SeaWiFS data were downloaded from the SeaWiFS Program [McClain et al., 1998] and processed at Washington University. In the first stage of processing, the scattering by air molecules was removed from the total reflectance using a Rayleigh correction procedure developed by [Vermote and Tanre, 1992], which also included nominal corrections for ozone and water vapor absorption. Next, the pixel radiance values were transformed to reflectance. The calculated spectral reflectance values (fraction of radiation reflected) represented the combined reflectance from the land, water clouds and the ambient aerosol. The LAC data from adjacent spacecraft swaths were then merged and georeferenced to produce contiguous coverage for Eastern North America shown in Figures 2a-d. Routine quantitative techniques for the retrieval of aerosol optical depth over land from satellite sensors are not yet available, but it is an area of active research [King et al., 1999]. For this reason, the SeaWiFS satellite data were used here primarily for detecting the change in surface albedo due to the smoke aerosol.
The smoke aerosol was also detected using visual range data at over 200 surface meteorological stations distributed throughout Eastern North America. The visibility data were appropriately filtered for precipitation, corrected for relative humidity, and converted to the light extinction coefficient using the Koshmieder equation and a Koshmieder coefficient of 1.9 [Husar et al., 1981]. The visibility data provided an estimate of the surface spatial and temporal distribution of the smoke plumes over the U.S. In addition, the long-term visibility data, since 1982, provided a historical context for the May 1998 Central American fires. The hourly National Weather Service surface observations were downloaded from the National Climatic Data Center (http://www.nws.noaa.gov/)
The chemical characteristics of the smoke aerosol were derived from the Interagency Monitoring of Protected Visual Environments (IMPROVE) aerosol monitoring network [Malm et al., 1994] (http://www2.nature.nps.gov/ard/impr/reporting.html). For the network of over 20 sites over the Eastern U.S., size-segregated PM10 and PM2.5 samples were collected twice a week, and the PM2.5 was chemically analyzed. The aerosol chemistry data helped characterize the chemical composition of the smoke aerosol during the passage of the smoke plume.
The regulatory PM10 network maintained by EPA for the purposes of the PM10 standard also provided limited aerosol data set on the surface concentration of the smoke. At several monitoring sites both PM10 and PM2.5 concentrations were available during the May 1998 smoke event. In addition, a number of states have begun to implement TEOM PM10 samplers collecting hourly PM10 data. These data were obtained from EPA's AIRS database for a number of cities throughout the Eastern US.
As the smoke plume passed over Eastern North America, the impact on the atmospheric radiation was measured by Multifilter Rotating Shadowband Radiometers [Harrison et. al., 1994], which measured the direct/diffuse solar radiation as part of the US Department of Agriculture Ultraviolet -B (USDA UVB) monitoring network (http://uvb.nrel.colostate.edu/UVB/home_page.html) [Gibson, 1991]. The vertical optical depth was measured by sun photometers in the Aerosol Robotic Network AERONET network (http://spamer.gsfc.nasa.gov) [Holben et al., 1998]. The characteristic size of the smoke aerosol was also estimated from the spectral vertical optical thickness data.
The possible impact of the Central American smoke plume on the ozone concentration throughout Eastern North America was detected by over 500 ozone monitors that are part of the U.S. Environmental Protection Agency's Aerometric Information Retrieval System (EPA AIRS) (http://www.epa.gov/airs/airs2.html) regulatory ozone monitoring system. The ozone data were used to examine whether the smoke plume contributed to the production of excess ozone on the ground.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Based on the TOMS and SeaWiFS data, it was concluded that smoke lingered over southern Mexico, Guatemala and Honduras and adjacent oceans throughout the 1998 spring season. The fires intensified in May, 1998 as shown in Figure 1, a series of daily maps of the smoke plume between May 4-28, 1998, based on the TOMS aerosol absorbing index [Herman et al., 1997].
A more detailed spatial distribution of the smoke is obtained from the 1 km resolution SeaWiFS satellite images. Figure 2a-d depict the daily smoke pattern for Eastern North America on May 14-16. The daily maps represent the spatial superposition of three complimentary aerosol signals; SeaWiFS, TOMS, and extinction coefficient. The color SeaWiFS images were used as a substrate for the numerical aerosol data derived from the TOMS sensor (green lines) and the extinction coefficient derived from surface-based meteorological network (red lines). The spectral reflectance images from SeaWiFS provide a rich visual context relevant to the smoke, including surface reflectance, position of cloud systems relative to the smoke, and indications of wind direction.
Figure 2a shows the SeaWiFS data, rendered on top of a
three-dimensional elevation map. The smoke dispersion plumes over Central
America appear mostly over low elevation terrain, while high elevation regions
are largely smoke free. This suggests
that the high elevation terrain act as a barrier to the dispersion of smoke.
Also, the vertical distribution of the smoke is less than the mountain peaks of
2 km. The daily fire map derived from the DMSP sensor (Figure 2a) shows that on
a specific day (May 15, 1998) fire spots occur at thousands of locations
throughout the Central America.
However, based on the currently available satellite “fire products” it
is not possible to estimate the magnitude of the smoke emissions.
The yellow color of the smoke (i.e. enhanced reflectance in the red compared to the blue wavelengths) allows the differentiation between clouds and smoke plumes. The spectral reflectance and yellow coloration of the smoke plumes near the source is discussed further in the section on smoke aerosol characteristics (section 6).
Throughout a ten-day period, May 7-17, 1998 (Figure 1), two major smoke transport episodes have occurred covering major fraction of continental Eastern North America with smoke. The geneses of both transport episodes were under similar meteorological conditions: There was poor ventilation over Central America for several days resulting in an accumulation of smoke over the source area. Subsequently, the accumulated smoke was transported to Eastern North America.
The first smoke transport episode impacted the Gulf States on May 8 and 9, 1998. The smoke pall (Figure 1) entered the USA through the Gulf Coast at Texas and turned eastward blanketing Louisiana, Arkansas, Alabama, Tennessee, Georgia and northern Florida with smoke. The entire transport event was confined to about one day.
The focus of this paper is on the second episode, which began with accumulation over Central America on May 10, 11, and 12, 1998. By May 12, a remarkably large (about 1000 miles diameter) smoke cloud covered virtually the entire Gulf of Mexico (Figure 1). On May13, the smoke plume began a swift journey northward through Texas along the Mississippi Valley.
By May 14-15 (Figure 2b and c), the smoke plume passed through Missouri and reached Wisconsin with the leading edge entering western Ontario. As a result, a visible 4,000 km long ribbon of smoke was created stretching from Central America to the Hudson Bay in Canada. Over the next two days the north-south smoke plume was shifted eastward by a westerly cold front, which resulted in a remarkable contrast of haziness (smokiness) ahead and behind the front (Figure 2c).
By May 16, the smoke reached Ohio, Pennsylvania, and West Virginia, while the remainder of Eastern North America was wiped clean of the smoke, by the passage of the advancing cold front. On all three days the SeaWiFS, TOMS, and extinction data show virtually identical spatial distribution for the advancing smoke plumes.
During the three-day smoke passage (Figures 2b-d) the contours of the smoke plume derived from the TOMS data correspond closely to the hazy smoke plume derived from the SeaWiFS data. The visibility-derived extinction coefficient, (red lines) also broadly correspond to the SeaWiFS and TOMS aerosol pattern. However, the highest surface extinction coefficients appears further north closer to the leading edge of the plume, while the highest TOMS signal is closer to the source in Central America. The causes of these slight deviations in the two aerosol measures are not clear, but it could be due to a deeper smoke layer near the source.
The spatial correspondence of columnar and surface-based aerosol signals indicate that the bulk of the smoke plume was not separated from the surface, but transported largely within the planetary boundary layer. Unfortunately, more detailed vertical aerosol profiles are not available for this smoke event.
The gross features of smoke emission density throughout the spring season were estimated from the TOMS absorbing aerosol index data. The fires during the 1998 spring season were significantly more intense then during the years prior or after. Hence, the excess aerosol concentration during 1988 compared to 1999 provides a convenient way of isolating the role of the 1988 smoke. Figure 3 shows the 1998 excess TOMS Absorbing Aerosol Index (AAI) averaged over March, April, May 1998. By rendering the excess aerosol index, the contribution from dust and other interferences was eliminated.
It is reasonable to assume that on the average, excess smoke concentration hot spots (Figure 3) correspond to the most intense smoke emission regions in the spring of 1998. Using this criterion, the highest smoke emission in the spring of 1998 occurred on the border between Mexico and Guatemala. Additional intense smoke production was observed over the southwestern coast of Mexico. The quantitative relationship between the fire "hot spots from the DMSP and AVHRR sensors and the smoke emission density is not known.
Earlier research suggested that biomass burning can produce significant quantities of tropospheric ozone in addition to the visible smoke [e.g. Fishman et al., 1991; Andreae et al., 1996]. The impact of the smoke plume on the concentration of noontime surface ozone (Figure 4) was evaluated by mapping the daily maximum ozone from over 500 eastern US ozone monitoring sites. Figure 4 also includes the outline of the smoke plume for each day. The daily ozone maps indicate that during the passage of the smoke plume, large parts of the eastern US had surface noon ozone concentrations in excess of 100 ppb. However, comparing the spatial domain of the smoke aerosol plume with the high ozone suggests that within the smoke plume the surface ozone concentrations are depressed compared to the surface ozone values just east and west of the plume. Hence, the smoke plume did not appear to add ozone to the existing values, but rather actually depressed the ambient ozone values. This observation does not preclude the generation of excess ozone further along the path of the smoke plume [Fishman et al., 1991]. As shown below, the smoke aerosol reduced the solar radiation reaching the surface by 10-20%, particularly at shorter wavelengths near the photochemically active UV.
The national regulatory PM10 monitoring network operates by simultaneous sampling throughout the country every sixth day, on May10 and May 17. By an unfortunate coincidence, the smoke plume passed over Eastern North America between May14-16, so the 1000- station regulatory PM10 network could not provide spatial distribution data for the smoke event.
| 
|
| 
|
As the smoke pall stretched northward the hourly TEOM PM10 concentrations in St. Louis on May 15 (Figure 5), rose sharply from 20-60 mg/m3 to over 150 mg/m3. The 24-hour PM2.5 levels also increased from 27 mg/m3 on May 13 to 68 mg/m3 on May 15. The hourly PM10 concentration data at Austin, TX and Maryland Heights, WV also show a sharp concentration peak as the smoke plume passed over these locations on May 15-16, 1998. The duration of the smoke passage at these locations was generally less than a day and the highest daily average PM10 concentrations were generally below 100 mg/m3.
The most remarkable PM10 concentration data were measured in Brownsville, TX, where the peak hourly concentration exceeded 500 mg/m3 on May 8, 1998 (Figure 6a). For comparison, the EPA daily maximum National Ambient Air Quality Standard for PM10 is 140 mg/m3. Brownsville is located on the US-Mexican border on the Texas coast where the smoke incursions were most frequent (Figure 1).
The passage of the smoke plume across eastern Missouri and Illinois (Figure 6b) on May 15, 1998, was recorded by hourly surface visibility observations as well as by the hourly PM10 concentration in St. Louis, MO. The observations at five meteorological sites in the 500 km size region indicate the simultaneous rise and fall of the extinction coefficient as well as the PM10 concentration. This indicates that the smoke plume was spatially homogeneous on the surface and its characteristic plume size was greater than the 500 km size. As indicated in Figure 6b, the aerosol extinction to mass ratio of the smoke plume was about 4 m2/g, which corresponds to the literature values for smoke [Reid et al., 1998].
| 
|
Several in-situ and remote sensing instruments captured the physical, chemical, and optical properties of the May, 1998 Central American smoke plumes.
Size segregated PM10 and PM2.5 concentrations measured at the IMPROVE [Malm et al., 1994] monitoring site at Big Bend, TX indicate that during the smoke event over 75% of the PM10 mass was in the PM2.5 size range. Figure 7 shows daily average nephelometer measured light scattering coefficient for May 1998, as well as the fine particle mass concentration sampled twice a week at the IMPROVE monitoring site at Big Bend, TX. Also, the fine particle mass concentration was apportioned into five aerosol components, sulfate, organics, elemental carbon (LAC), soil, and other. Here again, aerosol extinction and fine mass closely track each other through out the month with aerosol extinction to mass ratio of about 4 m2/g. This indicates a rather stable size distribution throughout the month. The magnitude of the aerosol extinction and mass were rather modest (<50 mg/m3) compared to Brownsville, TX, Austin, TX and St. Louis, MO. This may be due to the fact that Big Bend was on the edge of the smoke plume throughout the smoke season (Figure 1) and also due to its elevation of 1,030 m.
Examination of SeaWiFS and TOMS data showed that the smoke plume impacted Big Bend Texas on May 26-27, 1998. The chemical composition data show that during the smoke passage on May 27 the concentration of organics (50%), elemental carbon (2%), sulfate (26%), soil (3%) and "other" (19%) species increased significantly. The organics and elemental carbon are known tracers of biomass burning. However, the sharp rise in sulfate to 10 mg/m3 on May 27, 1998 eludes a plausible explanation.
Figure 7. Daily average light scattering coefficient and chemically speciated fine mass concentration for the IMPROVE monitoring site at Big Bend, TX, May 1998.
The spectral optical thickness data derived from the AERONET network [Holben et al., 1998] at Bondville, IL on May 15, 1998 (Figure 8) show that the vertical optical depth of the smoke column reached t=4 at 0.55 mm wavelength. Furthermore, the AERONET spectral aerosol optical thickness data show an unusual peak at 0.5 mm. This shape of the spectral optical depth is markedly different from the more traditional power-law spectral extinction with Angstrom exponent of a=1.5 that is seen prior and after the passage of the smoke event. The best fit Angstrom exponent of a=0.6 during the smoke event, but clearly the curve is not of the power-law type. Rather the best fit theoretical curve is obtained by assuming an aerosol size distribution with volume mean diameter of 0.6 mm, and refractive index m=1.5-0.01i.
 
|
The impact of the smoke on the solar radiation reaching the surface was evaluated using shadowband radiometer data [Harrison et al., 1994] at Big Bend, TX (Figure 9). The smoke impact was estimated from the difference in the surface radiation during the clear day, June 19, 1998, and a smoky day, May 27, 1998. On the clear day, the total radiation reaching the surface at l=0.5 mm exhibits a smooth sinusoidal diurnal pattern. On a hazy day, the total radiation reaching the surface is similar but the amount is less, and the diurnal curve shows ripples superimposed on the main sinusoidal pattern. The fraction of the total radiation that is lost due to smoke aerosol is displayed in the top right inset of Figure 9. At 0.415 mm the lost fraction of the total radiation reaching the surface is 10-15%, compared to the clear day. On the other hand, the lost fraction at 0.865 mm is only 0-6%. It is interesting to note that the largest surface radiation loss occurs during morning and afternoon hours with a clear minimum at noon when the solar zenith angle is small. In addition, their airmass path length through the smoke plume is larger in the morning and evening compared to noon. This diurnal behavior can be attributed to the elongated aerosol phase function in the forward direction and reduced backscattering at noon. The significant reduction of downwelling of solar radiation and increased upwelling reflectance by the smoke aerosol is in stark contrast to the radiative effect of dust clouds. For similar aerosol optical depth, the dust aerosol has only 1-2% perturbation of the net radiative flux [Husar et al., 2000].
A major impact of the smoke aerosol is the re-distribution of direct solar radiation into diffuse skylight. The magnitude of direct normal solar radiation on the clear and smoky day is shown in the top left inset of Figure 9. On the clear day, June 19, 1998, the direct normal radiation rises and falls rapidly during the morning and evening hours, and remains roughly constant over 6 mid-day hours. On the smoky day, May 27, 1998, the rise of the direct solar radiation is delayed by 2-3 hours, and the afternoon decay occurs 2-3 hours earlier. In fact, even during the noon hours, the direct normal radiation on the smoky day is only 70% of the corresponding value at a clear day. Most of the lost direct radiation is redistributed by the smoke aerosol as diffuse skylight. In Figure 9, for example, the magnitude of the diffuse radiation falling on a horizontal surface is twice on the smoky day compared to the clear day.
The fraction of the solar radiation that is prevented from reaching the ground due to the smoke layer is either backscattered to space or absorbed within the atmosphere. The backscattered fraction of the radiation is detectable in the SeaWiFS satellite data, as shown in Figure 10a for location 39 13' North, 86 5' West. The impact of the smoke is evaluated again by comparing the spectral reflectance of vegetation and soil on clear and smoky day. On the clear day (May 17, 1998) the vegetation reflectance is only 2-3% in the visible range, except at 0.55 mm where it increases to about 10%. In the near infrared, the vegetation reflectance sharply raises to over 0.4. The reflectance of soil under clear sky conditions increases monotonically from below 0.1 at 0.4 mm to 0.3 at 0.9 mm.
The SeaWiFS data show that on a smoky day (May 16, 1998), the smoke aerosol distorted the reflectance spectra in a remarkable manner. At short wavelength near 0.4 mm, the effect of the smoke aerosol is to increase the reflectance from less than 0.1 to over 0.25 for both vegetation and soil. Hence, the smoke has increased the near-UV surface albedo by about 20%. On the other hand, the effect of the smoke in the near infrared is to reduce the reflectance of bright surfaces such as vegetation and soil.
It is beyond the scope of this paper to explore these radiative effects in more detail. It is simply noted that at the blue wavelength over dark surfaces backscattering prevails compared to extinction, while in the near infrared the opposite is true. Thus, in assessing the radiative impact of aerosols the spectral characteristics of the radiative perturbation demand special attention.
Another feature of the radiative impact of aerosols is the discoloration of surfaces as seen through a smoke layer. For instance, under clear sky condition the spectral features of vegetation, i.e. color is rather pronounced, while on a smoky day the spectral reflectance is rather smooth and flat, devoid of spectral features. In visibility research this haze effect is referred to as discoloration and reduction of contrast.
The SeaWiFS satellite data provide further documentation of the unique optical properties of the smoke aerosol. Figure 10b illustrates the spectral reflectance of the smoke aerosol, sampled at various distances away from the source. To aid comparison, all the reflectance values were normalized to the reflectance at l=0.5 mm.
The reflectance spectra of aged smoke aerosol was determined at three locations: over the waters west of Acapulco, Mexico, west of Florida, and over Lake Ontario. The normalized aerosol reflectance spectra over the water were determined by subtracting the haze-free ocean reflectance from the total reflectance during the smoke period. The aerosol backscattering (reflectance) spectra near Acapulco and Florida was similar to the smoke near the fires: it exhibited the characteristic reflectance peak at 0.5-0.6 mm. The increase in reflectance between the blue and red wavelength is sufficiently large (up to 50%) such that the smoke plume appears as distinctly yellow in color as seen by the sensor satellite. Both of these locations were about 1,000 km from the source and most of the smoke transport occurred over the ocean. On the other hand, the reflectance spectra over Lake Ontario (3,000 km from the source) exhibited monotonic decrease from blue to red. These observations suggest that during the multi-day transit the size of the smoke aerosol is reduced with age as the smoke passed over the moist land surfaces. It is also worth noting that Central American smoke plume was embedded in a moist, cloudy airmass as it was transported across Eastern North America. In future research, it would be interesting to explore whether the peak of spectral reflectance at 0.6 mm is primarily due to backscattering or it is also influenced by preferential absorption near 0.4 mm.
The smoke caused exceedances of the PM standard, health alerts, and impairment of air traffic, as well as major reductions of visual range, and red sunsets. It was a major air pollution event covered by the research community as well as by the local and national media. At this stage a full assessment of the multiple effects of smoke is not possible. However, it is known that the Texas Natural Resources Office has carefully monitored smoke and before the main smoke pall arrived, the TNRO issued a health alert for the coastal region on May 12, 1998, and extended it to May 18. The fine particle concentrations in Texas exceeded 200 mg/m3 at several monitoring sites. The health complaints prompted public health researchers to initiate ad hoc quantitative measurements on the respiratory effects of the persisting smoke. The St. Louis media as well as elsewhere, have reported numerous citizen complaints on poor air quality. At the St. Louis Lambert Airport one of the runways was closed resulting in 30 % reduction in landing rates and significant flight delays on May 15, 1998.
This paper was prepared by an ad-hoc 'virtual' group of investigators consisting of data providers who shared their data through the Internet and investigators who used the shared data for analysis. A significant part of this study was to collaboratively integrate and fuse these multi-sensory data and derive a better understanding of the smoke event. The collaboration of the 'virtual' group was also mainly conducted through the Internet.
ARL, Air Resource Laboratory, Central American Forest Fires Smoke Prediction Using Hysplit_4. http://www.arl.noaa.gov/ready/yucatanfire.html, 1998
Andreae, M.O., J. Fishman, and J. Lindesay, The Southern Tropical Atlantic Region Experiment (STARE) Transport and Atmospheric Chemistry near the Equator-Atlantic (TRACE) and Southern African Fire-Atmospheric Research Initiative (SAFARI): An introduction. J. Geophys. Res., 101, D19, 23519-23520, 1996.
Artaxo,P., E. T. Fernandes, J.V. Martins, M. A. Yamasoe, P. V. Hobbs, W. Maenhaut, K. M . Longo, and A. Castanho, Large-scale aerosol source apportionment in Amazonia, J. Geophys. Res., 103, D24, 31837-31847, 1998.
Barnes, R. A., R. E. Eplee, Jr., F. S. Patt, and C. R. McClain, Changes in the radiometric sensitivity of SeaWiFS, Appl. Opt., 38(21), 4649-4664, 1999.
Christopher, S. A., J. Chou, J. Zhang, X. Li and R. M. Welch, Shortwave direct radiative forcing of biomass burning aerosols estimated from VIRS and CERES. Geophys. Res. Letters, (accepted for publication), 2000.
Elvidge, C. D., H. W. Kroehl, E. A. Kihn, K. E. Baugh, E. R. Davis, W. M. Hao, Algorithm for the retrieval of fire pixels from DMSP Operational Linescan System. In: Global Biomass Burning, J.S. Levine (Ed), The MIT Press, Cambridge, Massachusetts p.73-85, 1996.
Fishman J., K. Fakhruzzman, B. Cros, D. Nganga, Identification of widespread pollution in southern hemisphere deduced from satellite anlyses, Science, 252, 1693-1696, 1991.
Gibson, J.H., 1991. Justification and Criteria for the Monitoring of of Ultraviolet (UV) Radiation: Report of the UV-B Measurements Workshop. Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, C0 80523, p21.
Harrison, L., J. Michalsky and J. Berndt, Automated Multi-Filter Rotating Shadowband Radiometer: An Instrument for Optical Depth and Radiation Measurements, Appl. Opt., 33, 5118-5125, 1994.
Herman, J. R., P. K. Bhartia, O. Torres, C. Hsu, C. Seftor, E. Celarier, Global distribution of UV-absorbing aerosols from Nimbus 7/TOMS data. J. Geophys. Res, 102, D14, 16911-16922, 1997.
Holben, B N., T. F. Eck, I. Slutsker, D. Tanre, J.P. Buis, A. Setzer, E. Vermote, J. A. Reagan, Y. J. Kaufman, T. Nakajima, F. Lavenu, I. Jankowiak, and A. Smirnov, AERONET - A federated instrument network and data archive for aerosol characterization, Remote Sensing of Environment, 66, 1-16, 1998.
Husar R. B., J. M. Holloway, D. E. Patterson, W. E. Wilson, Spatial and temporal pattern of eastern U.S. haziness: A summary. Atmos. Environ. 15, 1919-1928, 1981.
Husar, R. B., J. M. Prospero, and L. L. Stowe, Characterization of tropospheric aerosols over the oceans with the NOAA Advanced Very High Resolution Radiometer Optical Thickness Operational Product. J. Geophys. Res. 102, 16889-16909, 1997.
Husar, R.B., D. M. Tratt, B. A. Schichtel, S. R. Falke, F. Li D. Jaffe, S. Gassó, T. Gill, N. S. Laulainen, F. Lu. M Reheis, Y. Chun, D. Westphal, B. N. Holben, C. Geymard, I. McKendry, N. Kuring, G. C. Feldman, C. McClain, R. J. Frouin, J. Merrill, D. DuBois, F. Vignola, T. Murayama, S. Nickovic, W. E. Wilson, K. Sassen, N. Sugimoto, The Asian Dust Events of April 1998, J. Geophys. Res. (submitted for publication), 2000
Lyons, W. A., T. E. Nelson, E. R. Williams, J. A. Cramer, T. R. Turner, Enhanced positive cloud-to-ground lightning in thunderstorms ingesting smoke from fires. Science, 282, 77-80, 1998.
Kaufman, Y. J, P. V. Hobbs, J. H. Kirchoff, P. Artaxo, L. A. Remer, B. N. Holben, M. D. King, D. E. Ward, E. M. Prins, K. M. Longo, L. F. Mattos, C. A. Nobre, J. D. Spinhirne, Q. Ji, A. M. Thomson, J. F. Gleason, S. A. Christopjer, and S. -C. Tsay, Smoke, clouds and radiation-Brazil (SCAR-B) experiment, J. Geophys. Res., 103, D24, 31703-31808, 1998.
Malm, W. C., J. F. Sisler, D. Huffman, R. A. Eldred, and T. A Cahill, Spatial and seasonal trends in particle concentration and optical extinction in the US, J. Geophys. Res., 99, 1347-1370, 1994.
McClain, C. R., M. L. Cleave, G. C. Feldman, W. W. Gregg, S. B. Hooker, and N. Kuring, Science quality SeaWiFS data for global biospheric research, Sea Technol., 39, 10-16 1998.
Parmenter, F. C. Picture off the month: Smoke from slash burning operations, Mon. Weather Rev., 99, 684-685, 1971.
Radojevic, M. and H. Hassan, Air quality in Brunei Darussalam during the 1998 haze episode, Atmos. Environ., 33, 3651-3658, 1999.
Reid J. S. and P. V. Hobbs, R. J. Ferek, D. R. Blake, J Vanderlei Martins, M. R. Dunlap, and C. Liousse. Physical, chemical and optical properties of regional hazes dominated by smoke in Brazil, J. Geophys. Res., 103, D24, 32059-32080, 1998.
Vermote, E. and D. Tanre, Analytical expression for radiative properties of planar Rayleigh scattering media, including polarization. J. Quant. Spectrosc. Radiat. Transfer, 47(4), 305-314, 1992.