Satellite Measurements of Atmospheric Aerosols
Rudolf B. Husar
Washington University, St. Louis
March 23, 2009
Satellite Measurements of Atmospheric Aerosols .......................................................................................... 1
Introduction ............................................................................................................................................... 1
Aerosol Physical and Optical Properties.................................................................................................... 4
Satellite Orbits ..........................................................................................................................................11
Principles of Satellite Aerosol Detection ..................................................................................................13
Challenges of Satellite Aerosol Remote Sensing ......................................................................................18
Aerosol Retrievals: Past, Present and Future ............................................................................................21
Satellite Information Systems ...................................................................................................................25
Applications ..............................................................................................................................................26
References: ...............................................................................................................................................30
Introduction
For the past decade Earth orbiting satellites sensors have provided near-real time,
kilometer resolution data on the global distribution of atmospheric aerosols. The satellites
have elucidated major pollution, dust and smoke events in stunning detail because
atmospheric aerosols are exceptionally well suited for satellite remote sensing. The sun
provides a stable light source and the light scattered by aerosols is easily detectable by
high resolution, spectral radiation sensors. These unique Earth observations have shown
great potential to aid Earth science and also to support societal decision making in many
areas, including air quality management, the detection and forecasting of major dust and
smoke events and the development of aerosol emission inventories. The global-scale
aerosol monitoring has the further potential to improve our understanding of aerosolinduced climate effects. Given the rich supply and the strong demand for satellite aerosol
data products, it is expected that these aerosol measurement systems will continue to
expand their utility. However, these high expectations should be tempered by the fact that
1
this measurement technique is among the most challenging and complex aerosol
measurements. The tools and methods are still maturing and the aerosol retrieval has
inherent physical and mathematical limitations that are yet to be understood and
specified.
The science of atmospheric aerosol optics is based on well understood fundamental
physical principles: Mie theory of light scattering and the theory of radiative energy
transfer through scattering and absorbing media. The very same physical laws that govern
satellite aerosol detection are applied to atmospheric visibility studies. In fact, satellite
aerosol detection and day-time vision by the human eye operate on the same general
principle of passive remote sensing (Figure 1): The radiation source is the sun; the target
is the Earth’s surface; the detector is an array of color sensitive sensors, may that be the
retina in the human eye or the satellite radiation sensor; in between the reflecting surface
and the light sensor is the radiatively active atmosphere consisting of aerosols, air, and
clouds.
Figure 1. Radiation detected by a satellite aerosol sensor
However, the two fields have different objectives and approaches and evolved rather
independently from each other. The objective of visibility studies is to measure and
model the ambient aerosol microphysics and chemistry and to evaluate the resulting
effects on the optical environment. The optical effects themselves can be computed
2
directly from the measured microphysical aerosol properties. Satellite aerosol remote
sensing on the other hand intends to solve the much more difficult inverse problem of
deriving the properties of atmospheric aerosols from the measured optical effects at the
top of the atmosphere.
The inherent problem with this kind of inversion is that the retrieval of the unknown
aerosol properties requires knowledge of those same properties which are not available.
The current practice for resolving the inversion paradox is to prepare a discrete set of
‘aerosol models’, and to choose the model that best fits the satellite optical data. Much of
the uncertainty and divergence in the current satellite aerosol retrievals arises from the
inability to find suitable aerosol models that match the observed radiances.
Much of the knowledge on ‘aerosol models’, including aerosol size distribution
dynamics, chemical composition and optical properties has been generated through
atmospheric aerosol and visibility research and it is directly applicable to the
improvement of satellite aerosol detection. Conversely, satellite aerosol monitoring can
significantly enhance the understanding of atmospheric aerosols, including emissions,
transport and spatio-temporal pattern. The art and science of satellite aerosol remote
sensing consists of combining the knowledge and insights from the complementary
research areas of atmospheric aerosols and of remote sensing.
The intent of this brief chapter is to bring together the recent developments on aerosol
microphysics and optics contributed by atmospheric aerosol and visibility research and
the developments in satellite remote sensing. For more details on atmospheric aerosols
the reader is guided to the excellent textbooks by Friedlander (2000a) and Seinfeld and
Pandis (1998) and to the exhaustive review of visibility science by Watson (2002). For
more details on satellite aerosol measurement systems see the authoritative article by
King et al. (1999) and the extensive published literature on satellite aerosol sensing and
data products. The chapter begins with summary of aerosol optical parameters relevant
to satellite detection. The bulk of this chapter is devoted to the physical principles and the
challenges of satellite aerosol detection. Following a section on data dissemination
3
systems, the chapter is concluded with illustrative application examples in Earth science
and air quality management.
Aerosol Dimensions, Properties and Models
This section reviews key characteristics of atmospheric aerosols that are relevant to
satellite remote sensing. Emphasis is placed on the observed regularities in size
distribution, light scattering efficiency, angular scattering pattern and vertical
distribution.
A distinguishing characteristic of atmospheric aerosols is the high dimensionality. At any
location and time, aerosols are distributed over particle diameter, D, which spans at least
two orders of magnitude (0.1 and 10 μm) in the optically active subrange. Atmospheric
particles are also composed of discrete chemical species with composition, C. Each size
class for a given chemical species may have a unique shape, S. A subtle, but for aerosol
optics significant dimension is the nature of aerosol mixing, X. Some of the aerosol
species are mixed externally, while other species are mixed internally in a single particle
(Figure 3). Externally mixed particles do not interact physically or chemically and their
radiative effects are additive. The optical effect of internally mixed particles is much
different and less predictable (White, 1986). The consequence of internal mixing is that it
results in increased particle size without changing the number of particles.
Figure 3. Externally, internally and layered mixtures of atmospheric aerosols.
While the characterization of gaseous species requires only the physical dimensions of
space and time (x,y,z,t), the full characterization of an aerosol system requires four
4
additional dimensions (x, y, z, t, D, C, S, X). Satellite sensors have high spatial resolution
(x,y) but they detect the optical effect of a vertical column. Hence, a satellite sensor
detects an aggregate contribution which is an integral over five dimensions (z, D, C, S,
X). The deconvolution of this integral is a notoriously ill-posed mathematical problem
that does not possess a unique solution. In other words, the number of unknown aerosol
parameters is much larger than the known entities from optical measurements. This
inversion problem is made extra uncertain by the fact that the aerosol optical signal at the
top of the atmosphere is weak compared to the noise from the surface reflectance. A
strategy toward the solution of the inversion problem is to reduce the complexity of the
multi-dimensional aerosol model. Such reduction can be achieved by observing some
regularities of atmospheric aerosol behavior. In this section, we highlight some of those
regularities and their implications to aerosol retrieval.
The key aerosol microphysical parameters relevant to the optical detection of aerosols are
shown in Figure 2.
5
Figure 2. a. Aerosol physical properties relevant to optical detection; b. Aerosol mass
distribution; c. Light scattering and absorption per unit aerosol volume (or mass);
The aerosol volume (or mass) is typically distributed (Figure 2b) between the fine
particle mode in the diameter range 0.1 to 1 to 2 m range and the coarse particle mode
above . The multimodal nature of atmospheric aerosols constitutes a major simplification
in formulating aerosol size distribution. In a nutshell, the model introduced by Whitby et.
al. (1972), Whitby (1978) states that each aerosol mode has a characteristic size
distribution, chemical composition and optical properties. Implicit in the model is that
each mode can be dealt separately, since particles in each mode are governed by different
sources, transformation, transport, and removal processes and also associated with
different effects. According to Friedlander (2000b), the multimodal concept has
dominated our understanding of atmospheric aerosol dynamics for the last thirty years.
The modal concept has also been widely adopted as the standard size distribution model
by the aerosol remote sensing community (Remer and Kaufman, 1998; Kahn et al., 2001;
Dubovik et al., 2002) .
Coarse mode atmospheric particles tend to originate from primary emissions, e.g.
windblown dust, sea salt, fly ash, road dust and they are irregular in shape. Coarse
particles tend to be chemically stable, and do not interact with fine particles or with
themselves, i.e. they are externally mixed. The radiative impact of coarse particle
mixtures are additive. An exception to this rule is when ambient air is ingested into cloud
droplets and the result is internal mixing of all particles in the cloud droplets. Sea salt
particles may also grow at high humidity.
Fine mode particles (below 1-2 m) are mostly formed by combustion and condensation
processes, resulting in spherical liquid droplets. The size distribution of fine particles is
highly dynamic and responsive to environmental conditions such as gas particle
conversion, relative humidity and chemical interactions among particles. Over urban
industrial regions the fine particle mode is composed of sulfate, organic and nitrate
species, each having somewhat different size distribution (Seinfeld and Pandis, 1998).
6
Fine particles tend to be internally mixed, but may exhibit multiple mass modes. Sulfate
particles for instance, occur in two distinctly different mass modes depending whether the
formation mechanism is photochemical gas particle conversion (mass mode ≈ 0.2 μm) or
through liquid phase reactions (mass mode ≈ 0.5 μm) (McMurry et al., 1996). A further
problem arises from the hygroscopicity of fine particles. At high relative humidity,
hygroscopic sulfate particles tend to grow significantly in size, while hygrophobic smoke
particles will grow less. As a consequence, the optical properties of aerosols, including
the phase function and extinction coefficient, will also depend on the ambient relative
humidity of the air layer.
On the aggregate, fine particles in urban, industrial regions exhibit another regularity of
increased mass median diameter with increasing concentration. This is significant
because it allows the formulation of dynamic aerosol models that directly relate the
spectral and angular scattering characteristics to aerosol concentration (Remer and
Kaufman, 1998). This regularity is illustrated through the two independent set of
observations shown in Figure 4 a and b.
At low optical thickness the volume-mean
diameter near Washington D.C. (Figure 4a) was about 0.2 μm, but increased to about 0.5
μm at high AOT (Dubovik et al., 2000?).
7
Figure 4a. Size distribution of urban, industrial aerosol as a function of optical thickness
b. Change in the phase function as a function of extinction coefficient c. Phase function at
different global locations.
The most critical aerosol optical property for satellite aerosol retrieval is the angular
scattering1 since satellites detect the cumulative scattering of sunlight in the direction of
the sensor (Figure 4b). The directional dependence of scattering is defined by the phase
function P(θ) which is the ratio of the energy scattered into direction, θ, to the average
energy scattered into all directions. The scattering angle, θ, is the angle between the sun,
the aerosol volume and the sensor (Fig.4b). Unlike the symmetric Rayleigh scattering by
pure air, the scattering phase function for aerosols is elongated in the direction of the
incoming light.
Based on extensive monitoring over different locations in the USSR (Figure 4b),
Barteneva (1960), observed a remarkable dependency of the phase function shape on the
aerosol extinction coefficient. With increasing bext(m-1), the ‘cigar’ shaped aerosol phase
1
Recent analysis by Mishchenko et al. (2009) found a significant deviation between two popular satellite
aerosol products, MODIS and MISR. The deviation was systematic and depended on scattering angle. This
is an indication that the phase functions used for the two satellite aerosol products differ considerably. This
gives further justification for a closer examination of aerosol phase functions.
8
function becomes increasingly shifted toward the forward direction. This pattern is
consistent with the notion that at higher bext, the characteristic particle size increases and
the scattering is more and more forward directed. Since the humidity of the atmosphere,
is a key driver for particle growth, Barteneva’s data implies that for hygroscopic aerosols,
P also depends on the relative humidity of the ambient air. The broader applicability of
the Barteneva’s observations was confirmed through airborne polar nephelometer
measurements in the US by Johnson (1981) who concluded that the directional scattering
properties of the atmospheric aerosol can be deduced from a measurement of the total
scattering coefficient alone. Conversely, satellite measurement of directional scattering
can be related to the more generally applicable total scattering and vertical optical
thickness.
The shape of the scattering phase function for atmospheric aerosols collected over
different parts of the world (Sakunov et al., 1992) is shown in Figure 4c. It is evident that
the forward scattering part of the phase function in the 30-60o direction is virtually the
same for all measured aerosol types, continental, oceanic and dust aerosols. This is a
significant finding since it establishes the direct proportionality between the forward
scattering and total scattering regardless of the aerosol type. Satellite measurement of
forward scattering is only possible on geostationary platforms when the sensor is directed
toward the sun. The phase function at backscattering angles (90 < θ < 180), shows more
deviation between aerosol types. Continental aerosols at visual range over 10 km and
oceanic aerosols have similar values. On the other hand, Sahara dust and continental
aerosol at visual range below 10 km have a factor of 2-3 lower P.
Another regularity of atmospheric aerosols show that at any given region, there is a
remarkably high correlation between the total light scattering coefficient, bscat and the
mass concentration of fine particles below 2.5 μm. The slope of the correlation
corresponds to scattering efficiency, α ≈ 3-5 m2/g, which is generally applicable for
sulfates, nitrates and organics in the fine particle mode. For atmospheric dust particles,
the corresponding extinction efficiency per unit mass is much less (α ≈ 0.6 m2/g), as
documented by White (1986). The observed regularity and ubiquity of this relationship
9
implies that there is a convenient conversion from fine particle mass concentrations, Mf
(μg/m3), to the optically relevant bscat= α *Mf. In visibility management and research, this
relationship is used to determine the visibility degradation from aerosol mass
measurements (Malm and Hand, 2007). Conversely, this same relationship allows the
estimation of fine particle columnar mass concentration Mfc[g/m2] from measured optical
thickness Mfc=τ/ α ]. In both cases due consideration needs to be given weather the
aerosol extinction is due to coarse or fine particles.
A further regularity pertains to the aerosol layering. It is known that aerosols tend to
reside in distinct layers of the atmosphere, as illustrated schematically in Figure 5. Figure
5b and 5c also illustrate the layering and mixing processes in the atmosphere. The
vertical distribution of aerosols can be subdivided into three layers with each layer having
different content, aerosol lifetime and mixing characteristics. The stratospheric or Junge
layer (Fig. 5a) is composed primarily of volcanic sulfuric acid aerosols that circle the
Earth at 10-15 km elevation for a year or two after volcanic eruptions. Below the
stratosphere is the tropospheric layer which is the carrier of dust, smoke or occasional
industrial haze ejected from the boundary layer. Tropospheric aerosols are thin pancakelike layers that may cover 1000 km size areas and tend to circle the Earth during their 1-2
week lifetime. The individual layers of tropospheric dust, smoke and haze do not
interact, hence their optical effects are additive.
10
Figure 5. a. Schematic representation of the aerosol vertical distribution. b. Astronaut
photo of stratospheric aerosol layers. c. Complex physical and radiative interaction
between aerosols and clouds.
The bottom layer is the planetary boundary layer (PBL) that contains all the sources and
sinks of atmospheric aerosols (Fig. 5). Precipitation and other removal mechanisms
restricts the aerosol lifetime in the PBL to 3-5 days. The shorter lifetime is responsible for
the spatial and temporal aerosol patchiness in the PBL. The intense mixing is partly
responsible for internal mixing of most aerosols in the PBL. The PBL potentially contains
a large variety of aerosols from anthropogenic and natural sources.
Each aerosol type shown in Figure 5a has its own sources, transport and optical
characteristics and therefore deserves to be treated as a separate aerosol type with specific
aerosol model parameterization. The aerosol types that are in separate boxes are
externally mixed and their respective radiative effects are additive. A group of aerosol
types in the boundary layer that are in adjacent boxes represent components that are
generally internally mixed and their physical, chemical and optical properties and
behaviors need to be considered jointly.
The above brief summary of aerosol characteristics highlights some of the regularities of
aerosol systems. It also offers physically-based classification of aerosol types. Clearly,
the formulation of the specific sub-models for each type requires considerable joint effort
by both the aerosol and remote sensing specialists. The classification is also compatible
with chemical transport models, and allows a tighter coupling of Earth observation and
modeling systems.
Satellite Orbits
Current Earth-observing satellites are circulating the Earth either in geostationary or polar
orbit, representing two different aerosol sensing and monitoring opportunities.
Geostationary satellites orbit roughly in the plane of the Equator at about 36,000 km,
11
such that their rotational speed is identical to the Earth daily rotation (Figure 6a). This
fixed geostationary perspective allows the monitoring of dynamic meteorological
phenomena such as, major dust clouds, smoke plumes or regional pollution events as they
evolve throughout the day. Current geostationary meteorological sensors are semiquantitative indicators of aerosols since they carry a single broadband sensor in the
visible range. However, for the elucidation of fast changing smoke, dust, and other
aerosol events, the sequential data every 30 minutes provide a compelling visualization of
aerosol emissions and transport dynamics. Geostationary aerosol observations in the
‘twilight zones’ (Figure 6a) at dawn and dusk are particularly helpful in detecting very
thin elevated aerosol layers, but the large distance from the Earth limits the sensor’s
spatial resolution to about 1km. Data from geostationary meteorological satellites are
routinely distributed through the Internet, minutes after they are taken and are used in
forecasting and other decision support applications.
Polar orbiting platforms rotate around the Earth in a plane that passes through the two
poles (Figure 6a). The satellite elevation is adjusted to about 100 km above the Earth
surface which corresponds to a full rotation every 90 minutes. Due to the low orbit, each
consecutive rotation samples only a swath of the Earth surface that is at most 3000 km
wide. Each swath is incremented such that the polar orbit is sun-synchronous, i.e. the
sensor will always observe the ‘sunny side’ of the Earth at the same local time of the day.
However, polar orbit means that any location on Earth can be observed only once a day
or less. The low orbit permits sensors to have higher spatial resolution and geo-location
accuracy. Most quantitative aerosol datasets and climatologies are derived from polar
orbiting satellites carrying a variety of sensors.
Figure 6. a. Geostationary and polar satellite orbits. b. A-Train satellite constellation.
12
There is an array of sensors that are currently deployed for the characterization of
atmospheric aerosols. It is fair to say, however, that no sensor can by itself fully
characterize the complex, dynamic, multi-dimensional aspects of the global aerosol
system. An opportunity for the integration of multiple sensors is being pursued by the ‘ATrain’ (Figure 6b), a synchronized constellation of satellites that fly on near-identical
orbits as a pack (Stephens et al., 2002). The constellation provides near-simultaneous
observations of the same part of the Earth through dozens of sensors on eight satellites.
Principles of Satellite Aerosol Detection
Satellite aerosol measurement is an optical technique for detecting the light scattered by
atmospheric aerosols. Based on the radiative signal at the top of the atmosphere, certain
integral properties and spatio-temporal pattern of atmospheric aerosols are inferred. The
four components of satellite aerosol detection are shown schematically in Figure 1: (1)
The radiation source is the sun; (2) the target is the Earth’s surface; (3) the radiation
detector is on the satellite; (4) in between is the radiatively active atmosphere consisting
of aerosols, air and clouds. The solar radiation reaches the surface either as a direct beam
(A) or indirectly as diffuse (B) light scattered by the air and aerosol particles (Fig. 1). A
fraction is backscattered toward the space (C). This backscattered solar radiation is the
primary signal that is used for the satellite detection of atmospheric aerosols. Most of the
incoming radiation is absorbed at the surface but a fraction, R0, is reflected. The
upwelling radiation is again attenuated by the atmosphere, but also enhanced by aerosol
scattering into the sensors line of sight.
The radiative signal received by a satellite is the combined contribution of surface
reflectance and aerosol reflectance. For the sake of illustration, a single scatter radiative
transfer theory is used here to show the aerosol interactions with solar radiation. The
separation of the aerosol signal from the surface reflectance can be achieved by the
solution of the appropriate radiation transfer equation as derived in (Husar and White,
1976). The resulting relationship between the surface reflectance and aerosol optical
parameters is shown in Figure 7.
13
Figure 7. Modification of surface reflectance R0 by aerosol acting simultaneously as a
filter and a reflector.
The aerosol perturbation of the upwelling radiation consists of two competing effects. On
one hand, light scattering particles tend to add reflectance to the brightness of surfaces.
This contribution depends both on the aerosol optical thickness, τ, as well as the aerosol
scattering phase function, P. The other role of aerosols is to act as a filter which
exponentially diminishes the upwelling radiation from a reflecting surface and from
aerosol scattering itself. The net result of these two competing aerosol effects is
illustrated in Figure 8d, where the apparent reflectance is plotted against the aerosol
optical thickness. The bundle of curves corresponds to different values of surface
reflectance R0. The simplest case is when the surface is black, R0=0 and the received
radiation is purely due to aerosol scattering along the path, i.e. the path radiance PR= (1e-τ)P (Kaufman, 1993). For optically thin aerosol layer, τ<1, PR is proportional to the
optical thickness since self-extinction of the aerosol layer is minimal. As the aerosol layer
approaches τ>2, self extinction becomes more significant and by τ >4, the apparent
reflectance asymptotically approaches the value of P which is the case for radiative
equilibriμm. Good examples for the strong path radiance are dark surfaces such as ocean
14
or vegetation at blue wavelength where R0 is small (<0.05), compared to P≈0.3 (Figure
8b and 8c)
The opposite case occurs when the surface reflectance R0, is high compared to P. In this
case, the addition of aerosol optical thickness actually diminishes the initial high
reflectance of the surface since the filter term dominates over the aerosol reflectance
term. As the aerosol layer above the bright surface becomes optically thick, τ >4, the
apparent reflectance asymptotically approaches the value of aerosol reflectance, P. For
example, an aerosol layer on top of bright clouds will tend to decrease the brightness of
clouds. Figure 8a is an example when a layer of dust on top of white clouds reduces the
cloud reflectance.
The next interesting case is when the aerosol phase function has the same value as the
surface reflectance function, P= R0. in the direction of sensor’s line of sight. In this case
adding aerosol will not change the apparent reflectance. At P≈R0, e.g. soil and vegetation
at 0.8 μm, the reflectance is unchanged by typical haze aerosols and aerosol detection is
not possible. The critical measure, whether aerosol will increase or decrease the apparent
reflectance is the ratio of the aerosol and surface reflectances, P/R0.
15
Figure 8. Spectral reflectance with and without aerosol. a. Cloud b. Vegetation c. Ocean.
d. Apparent surface reflectance as a function of P/R0 and AOT.
Figure 8d also shows that with increasing aerosol optical thickness, the apparent
reflectance of bright and dark surfaces converge toward the value of the aerosol
scattering function, P. In other words, with increasing AOT the contrast between black
and white surfaces diminishes and they become indistinguishable by τ>4. In this case of
an optically thick aerosol layer, the filtering and scattering effects of aerosols are
balanced, i.e. the aerosol system reaches radiative equilibrium. The implication is that the
retrieval of τ for optically thick aerosol layers (τ>4) is not possible. Aerosol retrieval is
also limited when P≈R0, which occurs over the bright surfaces of deserts and rocky
terrains.
The optical thickness, τ of the aerosol column is derived from the excess reflectance, i.e.
the difference between the apparent reflectance of the top of the atmosphere. Through the
radiative transfer theory it can be shown that the excess reflectance and the aerosol
optical thickness are related by the following simple expression, τ =-ln(R-P)/(R0- P). The
16
value of P is obtained from fitting the observed and retrieved surface reflectance spectra.
For example in light haze at 412 nm it is found that P=0.38.
In order to visualize the non-linear radiative effects of aerosols, Figure 9 provides graphic
illustration of these effects. Figures 9a-d show the satellite-received apparent reflectance
at 412, 555, 670 and 870 nm wavelengths. Figure 9e is the measured surface reflectance,
R0 in the absence of aerosol and clouds. Figure 9f shows the pattern of AOD at 412 nm
retrieved from the apparent reflectance R in Figure 9a and surface reflectance R0.
17
Figure 9. a-d. Measured apparent reflectance at 412, 555, 670 and 870 nm wavelength. e.
Surface reflectance in the absence of aerosol and clouds. f. Derived aerosol optical Depth,
AOD.
At blue wavelength (412 nm) Figure 9a, the apparent reflectance, R, reaching the satellite
sensor is dominated by the haze since the surface reflectance is low over the ocean and
vegetation. The aerosol patterns are clearly discernable, while all the surface features are
obscured by haze. Hence, the blue wavelength is well suited for aerosol detection over
land (Hsu et al., 2006) but surface characterization is difficult (see Figure 8b). Over the
ocean the apparent reflectance in the blue is also dominated by the added aerosol
reflectance (see Figure 8c). At green wavelength (555 nm) Figure 9b, the apparent
reflectance is still dominated by haze. At red wavelength (677 nm) Figure 9c, the
apparent reflectance is contributed significantly by both surface reflectance and aerosol
reflectance. However, the magnitude of the surface reflectance differs greatly for
vegetated and solid surfaces, such as urban centers. Hence, at the red wavelength aerosol
detection over land is difficult because it is very sensitive to the type of underlying
surfaces. Over the ocean the aerosol reflectance is high compared to the water
reflectance, and therefore quantitative aerosol detection is possible. The near IR
wavelength (865 nm), Figure 9d is uniquely suitable for aerosol detection over the ocean
since the ocean reflectance is below 1%. Hence, virtually all the apparent reflectance is
due to aerosol scattering. Over land, the reflectance of both vegetated and solid surfaces
is high and comparable to the aerosol reflectance P, i.e., P/Ro ~ 1 and aerosol detection is
not possible.
Aerosol interaction with light is a wonderful topic that never ceases to offer intriguing,
new counter-intuitive phenomena. The satellites as eyes in the sky, have offered a way to
visualize and enjoy these phenomena as a colorful complement to somewhat gray and dry
radiative transfer theory (e.g. Chandrasekhar, 1960)
Challenges of Satellite Aerosol Remote Sensing
18
Practical satellite aerosol measurements also face a number of additional complications
and challenges. Clouds obscure the detection of both underlying surfaces and
atmospheric aerosols. The areas delineated by a ‘cloud mask’ are unavailable for aerosol
detection. Since large portions of the Tropics and also at Northern latitudes are cloudy
throughout the year, cloudiness is a major limitation of satellite remote sensing of both
aerosols and surfaces. The cloud edges are not easily discernable from aerosols,
particularly during humid, hazy conditions. Thin cirrus clouds are also difficult to
identify. Consequently, improper cloud masks are a source of error in satellite-derived
aerosol fields. Cloud shadows on the Earth’s surface also constitute a complication for
aerosol retrieval.
Air or Rayleigh scattering of blue light always contributes to the radiation detected by the
satellite sensor. The magnitude of air scattering can be corrected by rigorous calculations
based on the known optical properties of air molecules, the elevation of the surfaces and
by application of well established radiative transfer code, such as offered by (Vermote et
al., 2002). Absorption by stratospheric ozone and tropospheric water vapor also attenuate
selected bands in the solar spectrum passing through the atmosphere. The stratospheric
ozone layer has an interfering absorption band in the visible range (0.52-0.74 μm).
Whenever possible, the satellite sensor bands are located in transmission windows, i.e. at
wavelengths away from major molecular absorption bands.
From the perspective of aerosol detection, the highly textured and colorful land reflection
from the Earth’s surface constitutes a significant background signal that needs to be dealt
with. The reflectance, i.e. the fraction of the incoming solar radiation reflected by a
surface, depends on many parameters most notably on the nature of the surface itself
(Fig. 10) and the geometric and spectral distribution of incoming and reflected radiation.
19
Figure 10. a. Reflection geometry for water, b. Reflection geometry for land and c.
Reflection geometry for vegetated surfaces.
Ocean surfaces reflect the incoming radiation, mostly by mirror-like, specular reflection,
Figure 10a. The areas of ‘sun glint’ can be identified from the sun-surface-sensor
geometry and can be masked out just like clouds (Khan et al., 2007). Ocean waves tend
to broaden the angular spread of the specular reflection and may also add spurious
reflectance from foaming white caps. Over land, the angular dependence of the reflected
radiation depends both on the angle of the incoming radiation as well as the angle of
reflected radiation and is defined by the Bidirectional Diffuse Reflectance Function
(BDRF) (Roujean et al., 1992). Solid land surfaces are mostly diffuse Labmertian
reflectors, with minor specular component (Fig. 10b). Vegetation is nearly diffuse
Labmertian reflector but exhibit an extra reflection ‘hot spot’ back toward the sun (Fig.
10c).
The BDRF is also strongly wavelength dependent which gives color to objects (Figure
11a). The spectral reflectance of water is low in the visible range and vanishes in the near
IR, making it practically black. Vegetation reflectance has a peak in the green (0.5 μm)
and rises sharply to over 40% in the near IR. Typical soil or rock reflectance gradually
increases from about 10-15% in the blue to over 40% in the near IR. The spectral
20
reflectance of many surfaces varies with season. For example, Figure 11b shows that the
seasonal variation of farmland reflectance at blue, green and red wavelengths exceeds a
factor of two.
Figure 11. a. Spectral reflectance of characteristic surfaces. b. Seasonal variation of
spectral reflectance for Illinois farmland
Unfortunately, the a-priory quantification of the spectral BDRF is not available for the
ever-changing Earth surfaces at all locations and all times. This necessitates the
extraction key features of BDRFs from the ‘dirty’ satellite images using spectral
reflectance values for the same pixel on different observation conditions (e.g. Raffusse,
2003).
Aerosol Retrievals: Past, Present and Future
Aerosol retrievals from atmospheric optical observations have a rich history, vigorous
present and bright prospects for the future. The abundance of atmospheric optical data
prompted early investigators to seek information about the aerosol properties via
inversion of the optical data. Following the Krakatoa eruption, Kiessling (1884)
21
concluded that the spectacular optical phenomena worldwide were caused by the volcanic
dust with characteristic size between 1 and 2 μm. Wegener (1911) used simple but most
elegant reasoning to infer that most of the atmospheric haze that scatters the visible light
is comparable in size to the wavelength of light. With the advent of computers, numerical
inversion procedures were developed to extract aerosol size distributions from spectral or
angular scattering data. For example, based on the inversion of spectral extinction data
Penndorf (1957) concluded that the peculiar blue sun and blue moon phenomena in
Europe were due to aged Canadian forest fire smoke with characteristic size of 0.6 to 0.8
μm. The most sophisticated and useful inversion scheme and service is offered by
Dubovik and King (2000) for the extraction of columnar aerosol size distribution data
from the AERONET (Holben et al., 1998) sun-sky photometer network.
Satellite aerosol detection begun in the mid-1960’s with the observation of ‘anomalous
gray shades’ in DMSP (Defense Meteorological Satellite Program) satellite images. By
the 1970’s, these gray shades were identified as Sahara dust (Carlson and Prospero,
1972), forest fire smoke over Central America (Parmenter, 1971), and other major
aerosol events. Quantitative aerosol detection begun in the early 1970’s with the
pioneering work of Griggs (1975) who observed that over dark water, the excess radiance
detected by a satellite over a southern California lake was correlated with the optical
depth (AOD) of the atmospheric aerosol column. Later Fraser (1976) pointed out that the
relationship between radiance and AOD depends on the type of aerosol, such as
windblown dust or industrial haze. Fraser applied formal radiative transfer theory, a new
derived relationship to estimate dust size distribution to the columnar mass concentration
for the Sahara dust. The early researchers including Mekler et al. (1977) and Kaufman
and Joseph (1982) used the LANDSAT data over water to make their satellite aerosol
measurements. The potential of geostationary satellites to augment surface-based
particulate air pollution measurements was illustrated by (Lyons and Husar, 1976). In
another seminal paper, Fraser et al. (1984), laid out a procedure for satellite-based
measurement of regional sulfate haze over the Eastern U.S., including the estimation of
the aerosol mass transport rates out of the region. By the 1980’s, a significant number of
satellites were deployed that yielded increasingly better aerosol characterization, and by
22
the 1990’s satellite Earth observations have produced multi-year global-scale aerosol
monitoring data.
Atmospheric aerosols constitute a significant nuisance to satellite remote sensing of the
Earth’s surface and of atmospheric composition. Much of the early work was therefore
focused on devising ‘aerosol correction’ algorithms and removing the spurious aerosol
signal. The earliest quantitative global scale aerosol measurements were obtained as byproducts arising from aerosol correction. Sea surface temperature detection by the
AVHRR sensor required a significant aerosol correction (Stowe, 1991) which yielded the
first multi-year global climatology of aerosol optical depth over the oceans (Husar et al.,
1997). The columnar ozone detection by the TOMS ozone sensor was also distorted by
absorbing aerosols and subsequently yielded the absorbing Aerosol Index, AI and
produced the first very useful global absorbing aerosol maps over land as well over
oceans (Herman et al., 1997). The ocean color and chlorophyll detection by the SeaWiFS
sensor (Gordon and Wang, 1994) also contributed significant advances through aerosol
corrections algorithms.
By the 1990’s a wide array of sensors that have been deployed for the characterization of
atmospheric aerosols. Some sensors provide specific aerosol characteristics. More
sophisticated satellite aerosol sensors seek to retrieve several aerosol properties including
spectral optical thickness and single scatter albedo from which one can infer the aerosol
type. These sensors use backscattered solar radiation in multiple narrow wavelength
channels (MODIS (King et al., 1992), SeaWiFS, (Gordon and Wang, 1994), polarization
of backscattered radiation (POLDER) (Deuzé et al., 2001), combined angular and
spectral scattering (MISR) (Diner et al., 1998). A common goal of the aerosol retrievals
is to extract the key properties of the aerosol system, usually the vertical aerosol optical
thickness (AOT) as a function of wavelength. The AOT is a measure of the total aerosol
burden and the spectral dependence is a clue on the size-composition of the aerosol
column. A common technique is to prepare a catalog of aerosol models for the surfaces
and to pick the appropriate model based on the best fit criteria. The aerosol models and
the surface models are typically tailored to the characteristics of the particular sensor. For
23
example, the two-wavelength AVHRR sensor can only accommodate a simple one
parameter model, while the multi-spectral MODIS sensor can, in principle, retrieve
multiple aerosol parameters. The most sophisticated MISR aerosol sensor detects both the
spectral and angular characteristics of the backscattered radiation and it uses many
aerosol and surface models from which a best fitting model is selected.
The performance of satellite aerosol sensors are evaluated primarily against observations
optical thickness data provided by the AERONET (Holben et al., 1998) sun photometer
network. The federated network has strategically located sites that characterize typical
aerosol regimes such as dusty desert locations, biomass smoke areas, marine background,
as well as urban industrial sites with complex aerosol mixtures. The validation results
appear to confirm the anticipated retrieval accuracy of MODIS and MISR products.
(Remer et al., 2008), (Abdou et al., 2005), (Kahn et al., 2005), (Kahn et al., 2007).
However, a direct comparison of MODIS and MISR aerosol data show significant
disagreements at the pixel level as well as between long-term and spatially averaged
aerosol properties, particularly over land (Liu and Mishchenko, 2008). Based on a longterm comparison of the Mishchenko et. al., (2009) conclude that the current satellitederived AOT and, especially aerosol microphysical characteristics remains
unsatisfactory.
It would be pretentious to speculate about the future development in satellite aerosol
remote sensing. However, it seems appropriate to express desirable developments from
the perspective of this reviewer.

Atmospheric aerosol systems are the result of dynamic processes that produce
particles with ever-changing size distribution, chemical composition and mixing
characteristics. In the future, it would be desirable to replace the static catalog of
aerosol models with dynamic models that are driven by basic physical and
chemical principles.

Land and ocean surfaces are not only highly textured spatially and spectrally, but
also change in time. More robust and continually updated surface reflectance
models would not only help aerosol retrieval but also contribute to Earth System
24
science and other benefit areas. In fact, a co-retrieval of aerosol and surface
reflectance would be ideal.

Future aerosol retrievals could utilize complementary and redundant information
from multiple sensors. For example, pixel level data fusion of MODIS and MISR
data could provide an improved aerosol model. Adding the TOMS/OMI data
would strengthen aerosol type selection, and fusion of geostationary and polar
satellite data could yield estimates on aerosol dynamics throughout the day.

Lastly, the concept of ‘aerosol retrieval’ could be expanded beyond the explicit
use of the optical aerosol signal processing. Processing of the spatial and temporal
signal could provide additional ways to deconvolute the integrated aerosol signal.
Satellite Information Systems
Earth observing satellites are by far the most prolific sources of environmental
monitoring data. Literally terabytes of digital data are collected, transmitted and
processed daily. An integral part of Earth observing satellite systems is the data flow and
processing network, i.e. the information system that collects and distributes the relevant
processed data to the end users. The first stage of the dataflow is from the satellite sensor
to the data silo where the processed primary data are deposited. This first stage, i.e. the
data acquisition stage, is specific to each satellite sensor and generally it is under the
control of the agency responsible for the satellite data acquisition and storage. Virtually
all the raw and processed satellite products are publicly accessible from their respective
agencies.
The second stage is the data usage stage. It incorporates the data flow from data archives
to the applications of the satellite data products in various societal benefit areas. This
stage involves considerable additional data processing, i.e. filtering, aggregation and
multi-sensory data fusion. Fortunately, data dissemination technologies offer outstanding
opportunities to fulfill the information needs of applications. The ‘terabytes’ of data from
satellite and other sensors can now be stored, processed and delivered in near real time.
Standardized computer-computer communication languages and Service-Oriented
25
Architectures (SOA) now facilitate the flexible processing of raw data into high-grade
‘actionable’ knowledge. Last but not least, the World Wide Web has opened the way to
generous sharing of data and tools leading to faster knowledge creation through
collaborative analysis and virtual workgroups. Recent developments in surface and
satellite sensing along with new information technologies now allow real-time, ‘just-intime’ data analysis for the characterization and partial explanation of the of major air
pollution events as well as more in-depth post-analysis (Husar and Poirot, 2005).
Nevertheless, users of satellite data still face significant hurdles. The new developments
introduced a new set of problems. The ‘data deluge’ problem is especially acute for
analysts interested in aerosol pollution, since the aerosol processes are inherently
complex, the numerous relevant data range from detailed surface-based chemical
measurements to extensive satellite remote sensing and the integration of these requires
the use of sophisticated data integration and modeling science methods. Unfortunately,
neither the current science, nor the tools and methods are adequate for the sophisticated
data integration. As a consequence satellite-derived, Earth observations are underutilized in making societal decisions.
A remedy is anticipated from the Global Earth Observation System of Systems (GEOSS),
(GEO, 2005) an emerging public information infrastructure for finding, accessing and
applying diverse data useful for decision makers. A motto of GEOSS is that ‘any single
dataset can serve many applications and any application requires many datasets’. GEOSS
is proactively pursuing the linking together of Earth observing systems so that a shared
virtual data pool can be combined and turned into actionable knowledge for many
societal benefit areas. A unique contribution of GEOSS is the adoption and promotion of
the advanced 'system of systems' (SoS) approach toward the integration of the
multiplicity of autonomous Earth observations and models.
Applications
26
There is increasing demand for satellite aerosol measurement in Earth science, air quality
management, and disaster management. Satellite aerosol observations are the only
available techniques to measure the distribution of aerosol over the entire globe. The
biogeochemical and climate applications of aerosols are based on the newly-derived
global aerosol data. Figure 12 shows the global distribution of aerosols in the month of
January and July measured by two satellite sensors. The MISR data, Figure 12 a,b are
representative of the total aerosol burden as measured by the vertical AOD. This includes
data from all aerosol layers, aerosol types, and aerosol size classes that contribute to the
AOD at 0.5 μm. The TOMS instrument (Figure 12 c,d) is sensitive to elevated aerosol
layers that are absorbing, such as wind-blown dust and biomass smoke. Non-absorbing
aerosols, e.g. sulfates and absorbing aerosols near the surface are not detected by TOMS.
The differences between the two sets of global aerosol maps are attributable to the
respective sensor sensitivities. The MISR July AOT shows the haze over the
anthropogenic aerosol sources areas of Eastern North America, East Asia and India, as
well as the dust and biomass smoke sources over Africa. The TOMS aerosol index only
highlights the source areas of dust and smoke. The integrated utilization of the available
multi-sensory satellite data for the characterization of global aerosols is work in progress.
Figure 12. Global aerosol maps for July and January; a.,b. MISR AOT; c.,d. TOMS
Aerosol Index
27
Satellite-derived global aerosol climatologies have improved our understanding of global
biogeochemical processes of sulfur, organics, mineral matter and other substances vital to
the Earth system. Satellite observations made it possible to derive aerosol source regions
(Prospero et al., 2002), transport pattern and indication of atmospheric lifetimes of these
substances.
With increasing concern about human-induced climate change, quantitative global-scale
monitoring of radiatively active atmospheric aerosols has become a major goal. On the
average, aerosol scattering tends to reduce the incoming solar radiation, while aerosol
absorption contributes to the heating of the atmosphere. The relative contribution of the
human induced aerosol burden on heating and cooling effects is not well understood.
Unfortunately, the current satellite-derived aerosol products are too uncertain to resolve
these subtle aerosol effects on climate (Mishchenko, 2009).
Air quality research has also benefited from satellite measurements, particularly for the
detection and quantification of natural and man-made air pollution events. Figure 13 is an
example of a smoke event which occurred in May 1998. Biomass smoke from forest and
agricultural fires in southern Mexico and Guatemala was transported across the entire
Eastern US and caused record surface aerosol concentrations through most of the
transport path. Figure 14a delineates the spatial pattern of bluish smoke plume and white
clouds on a May 15, 1998. The semi-quantitative absorbing aerosol index measured by
the TOMS sensor, Figure 13b, shows the daily pattern of the biomass smoke as it is
transported from Central America toward Eastern North America. Similar satellite
observations from the SeaWiFS and TOMS satellites have elucidated the transport of
wind-blown dust from the East Asian Gobi Desert to the West Coast of North America
(Husar et al., 2001).
28
Figure 13. a. Smoke plμme over Eastern U.S. (SEAWiFS), b. Daily TOMS pattern
Air quality management is a recent application area of satellite aerosol measurements. In
the past, air quality regulations for PM relied on surface-based monitoring networks to
enforce the air quality standards. More recently EPA’s Exceptional Event Rule (EE)
(Federal Register, 2008) explicitly encourages, the use of satellites for the documentation
of aerosol contributions which originate outside the jurisdiction of air quality control
agencies. As an example of an exceptional event, every spring, the rangeland grass in
Kansas-Oklahoma is burned by the cattle farmers resulting in major smoke plumes that
cover multi-state areas of the Midwest. The smoke event can be documents by the
combination of surface monitoring and satellite data. The ‘fire pixels’ in Figure 14 a
show the location of fires over Kansas derived from the MODIS satellite sensor, the noon
AOD derived from the SeaWiFS satellite sensor (Fig. 14 c,d) and the spatial pattern of
organic fine particle mass concentration derived from speciated aerosol data collected
through several surface monitoring networks (Fig. 14b). Based on such integrated set of
multi-sensory observations and associated analyses, the federal EPA has issued
exceedance waivers for sites that were impacted by the agricultural smoke. A more recent
application of satellite aerosol data is directed toward the estimation of the emission
inventories from ‘exceptional’ sources, such as forest fires, dust storms, and biogenic
organic aerosol sources. The improved emission inventories in turn are expected to
enhance the performance of air quality forecast models.
29
Figure 14. Multi-sensory characterization of agricultural smoke over the Midwest on
April 12, 2003: a. Fire pixels over Kansas from the MODIS sensor; b. Organic aerosol
concentration based on speciated chemical samplers; c. and d. Aerosol optical depth,
AOD derived from the SeaWiFS satellite.
Acknowledgements:
This work was partially supported by the NASA Grants NNL05AA00A and
NNX08BA33G as well as EPA Grant XA83228301-01-0. Erin Robinson and Janja Husar
were instrumental in the preparation of the manuscript and their help is greatly
appreciated.
References:
Abdou, W. A., D. J. Diner, J. V. Martonchik, C. J. Bruegge, R. A. Kahn, B. J. Gaitley, K. A.
Crean, L. A. Remer, and B. Holben. 2005. Comparison of coincident Multiangle Imaging
Spectroradiometer and Moderate Resolution Imaging Spectroradiometer aerosol optical
depths over land and ocean scenes containing Aerosol Robotic Network sites. J.
Geophys. Res 110, no. D10S07.
30
Barteneva, O. D., G. G. Sakunov, and A. A. Timerev. 1996. Spectral transparency of upper-level
clouds by ground-based measurements in various climatic zones.
Carlson, T. N., and J. M. Prospero. 1972. The Large-Scale Movement of Saharan Air Outbreaks
over the Northern Equatorial Atlantic. Journal of Applied Meteorology 11, no. 2: 283297.
Chandrasekhar, Subrahmanyan. 1950. Radiative transfer.
Deuzé, J. L., F. M. Bréon, C. Devaux, P. Goloub, M. Herman, B. Lafrance, F. Maignan, A.
Marchand, F. Nadal, and G. Perry. 2001. Remote sensing of aerosols over land surfaces
from POLDER-ADEOS-1 polarized measurements. Journal of Geophysical Research
106, no. D5.
Diner, D. J., J. C. Beckert, T. H. Reilly, C. J. Bruegge, J. E. Conel, R. A. Kahn, J. V. Martonchik,
et al. 1998. Multi-angle Imaging SpectroRadiometer (MISR) instrument description and
experiment overview. IEEE Transactions on Geoscience and Remote Sensing 36 (July 1):
1072-1087.
Dubovik, O., B. Holben, T. F. Eck, A. Smirnov, Y. J. Kaufman, M. D. King, D. Tanré, and I.
Slutsker. 2002. Variability of absorption and optical properties of key aerosol types
observed in worldwide locations. Journal of the Atmospheric Sciences 59, no. 3: 590-608.
Dubovik, O., and M. D. King. 2000. A flexible inversion algorithm for retrieval of aerosol
optical properties from Sun and sky radiance measurements. Journal of Geophysical
Research 105, no. D16.
Fraser, R. S. 1976. Satellite measurement of mass of Sahara dust in the atmosphere. Applied
Optics 15, no. 10: 2471-2479.
Fraser, R. S., Y. J. Kaufman, and R. L. Mahoney. 1984. Satellite measurements of aerosol mass
and transport. Atmospheric Environment (1967) 18, no. 12: 2577-2584.
Friedlander, S. K. 2000a. Smoke, Dust, and Haze: Fundamentals of Aerosol Dynamics. Oxford
University Press, USA.
Friedlander. 2000b. A brief history of aerosol dynamics. History of Aerosol Science” Verlag der
Oesterreichischen Akademie der Wissenschaften (Preining O. and Davis EJ Eds.),
Vienna.
GEO - Group on Earth Observations | Home. http://www.earthobservations.org/., 2005
Gordon, H. R., and M. Wang. 1994. Retrieval of water-leaving radiance and aerosol optical
thickness over the oceans with SeaWiFS: a preliminary algorithm. Applied Optics 33, no.
3: 443-452.
Griggs, M. 1975. Measurements of atmospheric aerosol optical thickness over water using
ERTS-1 data. Journal of the Air Pollution Control Association 25, no. 6: 622.
Herman, J. R., P. K. Bhartia, O. Torres, C. Hsu, C. Seftor, and E. Celarier. 1997. Global
distribution of UV-absorbing aerosols from Nimbus 7/TOMS data. Journal of
Geophysical Research 102, no. D14.
Holben, B. N., T. F. Eck, I. Slutsker, D. Tanre, J. P. Buis, A. Setzer, E. Vermote, J. A. Reagan,
Y. J. Kaufman, and T. Nakajima. 1998. AERONET-A Federated Instrument Network and
Data Archive for Aerosol Characterization-an overview. Remote Sensing of Environment
66, no. 1: 1-16.
Hsu, N. C., S. Tsay, M. D. King, and J. R. Herman. 2006. Deep blue retrievals of Asian aerosol
properties during ACE-Asia. IEEE Transactions on Geoscience and Remote Sensing 44,
no. 11: 3180.
31
Husar, R. B., and R. L. Poirot. 2005. DataFed and FASTNET: Tools for Agile Air Quality
Analysis. Environmental Manager (September): 39-41.
Husar, R. B., J. M. Prospero, and L. L. Stowe. 1997. Characterization of tropospheric aerosols
over the oceans with the NOAA advanced very high resolution radiometer optical
thickness operational product. Journal of Geophysical Research 102, no. D14.
Husar, R. B., and W. H. White. 1976. On the color of the Los Angeles smog. Atmospheric
Environment (1967) 10, no. 3: 199-204.
Husar, R. B., D. M. Tratt, B. Schichtel, S. R. Falke, F. Li, D. Jaffe, S. Gasso, T. Gill, N. S.
Laulainen, and F. Lu. 2001. Asian dust events of April 1998. Journal of Geophysical
Research 106, no. D16: 18317-18330.
Johnson, R. W. 1981. Daytime visibility and nephelometer measurements related to its
determination. Atmospheric Environment (1967) 15, no. 10-11: 1835-1845.
Kahn, R., P. Banerjee, and D. McDonald. 2001. Sensitivity of multiangle imaging to natural
mixtures of aerosols over ocean. Journal of Geophysical Research 106, no. D16.
Kahn, R. A., B. J. Gaitley, J. V. Martonchik, D. J. Diner, K. A. Crean, and B. Holben. 2005.
Multiangle Imaging Spectroradiometer (MISR) global aerosol optical depth validation
based on 2 years of coincident Aerosol Robotic Network (AERONET) observations.
Journal of Geophysical Research 110: D10S04.
Kahn, R. A., M. J. Garay, D. L. Nelson, K. K. Yau, M. A. Bull, B. J. Gaitley, J. V. Martonchik,
and R. C. Levy. 2007. Satellite-derived aerosol optical depth over dark water from MISR
and MODIS: Comparisons with AERONET and implications for climatological studies.
Journal of Geophysical Research-Atmospheres 112, no. D18: D18205.
Kaufman, Y. J. 1993. Aerosol optical thickness and atmospheric path radiance. Journal of
Geophysical Research 98, no. D2.
Kaufman, Y. J., and J. H. Joseph. 1982. Determination of surface albedos and aerosol extinction
characteristics from satellite imagery. Journal of Geophysical Research 87.
Kiessling, J., Meteorol. Zeitschr., 117 (1884).
King, M. D., Y. J. Kaufman, W. P. Menzel, D. Tanre, NGSF Center, and M. D. Greenbelt. 1992.
Remote sensing of cloud, aerosol, and water vapor properties fromthe moderate
resolution imaging spectrometer (MODIS). IEEE Transactions on Geoscience and
Remote Sensing 30, no. 1: 2-27.
King, M. D., Y. J. Kaufman, D. Tanré, and T. Nakajima. 1999. Remote Sensing of Tropospheric
Aerosols from Space: Past, Present, and Future. Bulletin of the American Meteorological
Society 80, no. 11.
Liu, Li, and Michael I. Mishchenko. 2008. Toward unified satellite climatology of aerosol
properties: Direct comparisons of advanced level 2 aerosol products. Journal of
Quantitative Spectroscopy and Radiative Transfer 109, no. 14 (September): 2376-2385.
doi:10.1016/j.jqsrt.2008.05.003.
Lyons, W. A., and R. B. Husar. 1976. SMS/GOES Visible Images Detect a Synoptic-Scale Air
Pollution Episode. Monthly Weather Review 104: 12.
Malm, William C., and Jenny L. Hand. 2007. An examination of the physical and optical
properties of aerosols collected in the IMPROVE program. Atmospheric Environment 41,
no. 16 (May): 3407-3427. doi:10.1016/j.atmosenv.2006.12.012.
McMurry, P. H., X. Zhang, and C. T. Lee. 1996. Issues in aerosol measurement for optics
assessments. Journal of Geophysical Research 101, no. D14.
32
Mekler, I., G. Ohring, and H. Quenzel. 1977. Relative atmospheric aerosol content from ERTS
observations. In COSPAR, Plenary Meeting.
Mishchenko, Michael I., Igor V. Geogdzhayev, Li Liu, Andrew A. Lacis, Brian Cairns, and
Larry D. Travis. 2009. Toward unified satellite climatology of aerosol properties: What
do fully compatible MODIS and MISR aerosol pixels tell us? Journal of Quantitative
Spectroscopy and Radiative Transfer 110, no. 6-7: 402-408.
doi:10.1016/j.jqsrt.2009.01.007.
Parmenter, FC. 1971. PICTURE OF THE MONTH: Smoke From Slash Burning Operations.
Monthly Weather Review 99, no. 9.
Penndorf, R. B. 1957. New Tables of total Mie scattering Coefficients for spherical particles of
real refractive indexes (1.33<«< 1.50)./. opt. Soc. Am 47: 1010.
Prospero, J. M., P. Ginoux, O. Torres, S. E. Nicholson, and T. E. Gill. 2002. Environmental
characterization of global sources of atmospheric soil dust identified with the NIMBUS 7
Total Ozone Mapping Spectrometer (TOMS) absorbing aerosol product. Reviews of
Geophysics 40, no. 1: 1002.
Raffuse, S. M. 2003. Estimation of daily surface reflectance over the United States from the
SEAWifS sensor. Washington University.
Remer, L. A., R. G. Kleidman, R. C. Levy, Y. J. Kaufman, D. Tanré, S. Mattoo, J. V. Martins, C.
Ichoku, I. Koren, and H. Yu. 2008. Global aerosol climatology from the MODIS satellite
sensors. Journal of Geophysical Research 113, no. D14.
Remer, Lorraine A., and Yoram J. Kaufman. 1998. Dynamic aerosol model: Urban/industrial
aerosol. Journal of Geophysical Research 103: 13859-13872.
Roujean, J. L., M. Leroy, and P. Y. Deschanps. 1992. A bidirectional reflectance model of the
Earth's surface for the correction of remote sensing data. Journal of Geophysical
Research 97, no. D18.
Rules: Revised Exceptional Event Data Flagging Submittal and Docμmentation Schedule to »
Federal Register [FR Doc E8-29747] [40 CFR Part 50].
http://www.thefederalregister.com/d.p/2008-12-16-E8-29747.
Sakunov, G. G., D. Barteneva, and V. F. Radionov. 1992. Experimental studies of light
scattering phase functions of the atmospheric surface layer. Izvestiya: 750.
Seinfeld, John H., and Spyros N. Pandis. 1998. Atmospheric chemistry and physics : from air
pollution to climate change. In . http://adsabs.harvard.edu/abs/1998acpa.conf.....S.
Stephens, G. L., D. G. Vane, R. J. Boain, G. G. Mace, K. Sassen, Z. Wang, A. J. Illingworth, E.
J. O’Connor, W. B. Rossow, and S. L. Durden. 2002. THE CLOUDSAT MISSION AND
THE A-TRAIN. Bulletin of the American Meteorological Society 83: 12.
Stowe, L. L. 1991. Cloud and aerosol products at NOAA/NESDIS. Global and Planetary
Change GPCHE 4, 4, no. 1/3.
Vermote, E. F., N. Z. El Saleous, and C. O. Justice. 2002. Atmospheric correction of MODIS
data in the visible to middle infrared: first results. Remote Sensing of Environment 83, no.
1: 97-111.
Watson, J. G. 2002. Visibility: science and regulation. J Air Waste Manag Assoc 52, no. 6: 628713.
Wegener, A. L., Thermodynamic der Atmosphere. Johann Ambrosius Barth, Leipzig (1911).
Whitby, K. T., R. B. Husar, and B. Y. H. Liu. 1972. The aerosol size distribution of Los Angeles
smog. Journal of Colloid and Interface Science 39, no. 1 (April): 177-204.
doi:10.1016/0021-9797(72)90153-1.
33
Whitby, K. T. 1978. The physical characteristics of sulfur aerosols. Atmospheric Environment
(1967) 12, no. 1-3: 135-159. doi:10.1016/0004-6981(78)90196-8.
White, Warren H. 1986. On the theoretical and empirical basis for apportioning extinction by
aerosols: A critical review. Atmospheric Environment (1967) 20, no. 9: 1659-1672.
doi:10.1016/0004-6981(86)90113-7.
34