Seasonal Variation of Aerosol Extinction Coefficients and...

12
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015 2669 ISSN: 2278 7798 All Rights Reserved © 2015 IJSETR AbstractIn this study, the horizontal and vertical distributions of aerosols over Bay of Bengal region are investigated for the latitudinal and longitude range of 0 o N -10 o N and 80 o E-95 o E respectively. Aerosol extinction coefficient profiles and aerosol optical depths (AOD) over Bay of Bengal region for one complete year 2012 are retrieved using Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) satellite data and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data respectively. A series of algorithms has been studied and understood to identify the aerosol and cloud layers and to retrieve the extinction caused by aerosol layers. The AOD over ocean at 555 nm is used for the study in order to link with the extinction profiles that are retrieved. The AOD values varied from 0.0726±0.022 in winter to 0.0952±0.0517 in spring. The maximum AOD’s are observed in summer (0.1394±0.0606) and the minimum are observed in autumn (0.0506±0.035). In general, cloud-cleared aerosol extinction profiles showed exponential decrease of aerosols with altitudes. In general, aerosols are concentrated below 1km MSL (Mean Sea Level). Above 1km, extinction profiles exhibited a secondary maximum of aerosol at 1.5 -2km above the surface in winter and spring. Horizontal upper air transport could have brought these elevated concentrations. The observed elevated layers in summer and autumn could be due to presence of clouds. Index TermsAOD, CALIPSO, MODIS, Retrieval Algorithms, Seasonal variation, Solution of lidar equation. I. INTRODUCTION Atmospheric aerosols are suspension of small liquid or solid particles [1], with radii varying from a few nm to larger than 100μm, in air. These particles can enter into the atmosphere either by natural sources such as volcanic ash formed during the eruption of a volcano, sea salt, desert dust etc., or by human activities such as vehicular emissions, burning of fossil fuels, emissions from chemical industries etc., The scattering and absorption of radiation by gas molecules and aerosols all contribute to the extinction of the solar and terrestrial radiation passing through the atmosphere. These aerosols and clouds have important impacts on the earth’s climate. Clouds reflect sunlight back to space and also absorb the outgoing terrestrial radiation. Therefore, the net effect of Manuscript received June, 2015. G. V. Raghava Rayudu, Department of Earth and Space Sciences, Indian Institute of Space Science and Technology, Hyderabad, India. clouds is determined based on the altitude of the cloud layer present in atmosphere. Aerosols also can scatter sunlight back to space, producing cooling at the earth’s surface. Whereas the absorbing aerosols such as black carbon particles warm the atmosphere, which can influence the atmospheric stability and suppress cloud formation which is a direct effect. Aerosols can also lead to the formation of clouds by cloud particles interacting with the cloud particles. Increase of concentration of aerosols leads to increase in the number of cloud particles formed on them. Hence, aerosols can also alter the cloud properties which is an indirect effect. There is still a lot of debate going on the study of aerosols and clouds, because the largest uncertainties in determining the future climate change are associated with the uncertainties in the distribution and properties of aerosols and clouds and their interactions. The satellite and ground based measurements prior to CALIPSO shows only the spatial distribution of aerosol layers but not at what altitude the aerosol layer is located, CALIPSO provides us this important piece of information. The Cloud -Aerosol Lidar and Infrared Pathfinder Satellite Observations was launched by NASA in collaboration with Centre National d’Études Spatiales (CNES), the French space agency, on April 28, 2006. It is a part of A-train satellite constellation orbiting the earth at 705 km polar sun synchronous orbit, the orbiting period is 98.5 minutes and has a 16 day repeat cycle. The orbit inclination is 98.2 o , thus providing a global coverage between 82 o N and 82 o S. Careful control of the CALIPSO orbit with respect to the other A-train satellites ensures the acquisition of collocated near- simultaneous measurements [2]. The primary instrument aboard on CALIPSO is Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP), which is an active sensor. Prior to launch of CALIPSO, aerosols are being observed from space using passive techniques relying on scattered sun light. Remote sensing of aerosol properties is difficult, however, owing to the great variability in aerosol sources, distribution and properties, effects of surface reflectance, and interference from clouds. Active sensing of aerosols with lidar offers unique capabilities which are complementary to those of passive techniques. Lidar is the only technique giving high resolution profiles of aerosols, and it is able to observe aerosol above bright surfaces, such as deserts and snow, and above bright clouds. Because lidar provides its Seasonal Variation of Aerosol Extinction Coefficients and Aerosol Optical Depths Over Bay of Bengal Region G. V. Raghava Rayudu

Transcript of Seasonal Variation of Aerosol Extinction Coefficients and...

Page 1: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2669

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

Abstract— In this study, the horizontal and vertical

distributions of aerosols over Bay of Bengal region are

investigated for the latitudinal and longitude range of 0oN

-10oN and 80

oE-95

oE respectively. Aerosol extinction

coefficient profiles and aerosol optical depths (AOD) over Bay

of Bengal region for one complete year 2012 are retrieved using

Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP)

satellite data and Moderate Resolution Imaging

Spectroradiometer (MODIS) satellite data respectively. A

series of algorithms has been studied and understood to

identify the aerosol and cloud layers and to retrieve the

extinction caused by aerosol layers. The AOD over ocean at 555

nm is used for the study in order to link with the extinction

profiles that are retrieved. The AOD values varied from

0.0726±0.022 in winter to 0.0952±0.0517 in spring. The

maximum AOD’s are observed in summer (0.1394±0.0606) and

the minimum are observed in autumn (0.0506±0.035). In

general, cloud-cleared aerosol extinction profiles showed

exponential decrease of aerosols with altitudes. In general,

aerosols are concentrated below 1km MSL (Mean Sea Level).

Above 1km, extinction profiles exhibited a secondary

maximum of aerosol at 1.5 -2km above the surface in winter

and spring. Horizontal upper air transport could have brought

these elevated concentrations. The observed elevated layers in

summer and autumn could be due to presence of clouds.

Index Terms— AOD, CALIPSO, MODIS, Retrieval

Algorithms, Seasonal variation, Solution of lidar equation.

I. INTRODUCTION

Atmospheric aerosols are suspension of small liquid or solid

particles [1], with radii varying from a few nm to larger than

100μm, in air. These particles can enter into the atmosphere

either by natural sources such as volcanic ash formed during

the eruption of a volcano, sea salt, desert dust etc., or by

human activities such as vehicular emissions, burning of

fossil fuels, emissions from chemical industries etc., The

scattering and absorption of radiation by gas molecules and

aerosols all contribute to the extinction of the solar and

terrestrial radiation passing through the atmosphere. These

aerosols and clouds have important impacts on the earth’s

climate. Clouds reflect sunlight back to space and also absorb

the outgoing terrestrial radiation. Therefore, the net effect of

Manuscript received June, 2015.

G. V. Raghava Rayudu, Department of Earth and Space Sciences,

Indian Institute of Space Science and Technology, Hyderabad, India.

clouds is determined based on the altitude of the cloud layer

present in atmosphere. Aerosols also can scatter sunlight

back to space, producing cooling at the earth’s surface.

Whereas the absorbing aerosols such as black carbon

particles warm the atmosphere, which can influence the

atmospheric stability and suppress cloud formation which is

a direct effect. Aerosols can also lead to the formation of

clouds by cloud particles interacting with the cloud particles.

Increase of concentration of aerosols leads to increase in the

number of cloud particles formed on them. Hence, aerosols

can also alter the cloud properties which is an indirect effect.

There is still a lot of debate going on the study of aerosols and

clouds, because the largest uncertainties in determining the

future climate change are associated with the uncertainties in

the distribution and properties of aerosols and clouds and

their interactions.

The satellite and ground based measurements prior to

CALIPSO shows only the spatial distribution of aerosol

layers but not at what altitude the aerosol layer is located,

CALIPSO provides us this important piece of information.

The Cloud -Aerosol Lidar and Infrared Pathfinder Satellite

Observations was launched by NASA in collaboration with

Centre National d’Études Spatiales (CNES), the French

space agency, on April 28, 2006. It is a part of A-train

satellite constellation orbiting the earth at 705 km polar sun

synchronous orbit, the orbiting period is 98.5 minutes and

has a 16 day repeat cycle. The orbit inclination is 98.2o, thus

providing a global coverage between 82oN and 82oS. Careful

control of the CALIPSO orbit with respect to the other

A-train satellites ensures the acquisition of collocated near-

simultaneous measurements [2].

The primary instrument aboard on CALIPSO is

Cloud-Aerosol Lidar with Orthogonal Polarization

(CALIOP), which is an active sensor.

Prior to launch of CALIPSO, aerosols are being observed

from space using passive techniques relying on scattered sun

light. Remote sensing of aerosol properties is difficult,

however, owing to the great variability in aerosol sources,

distribution and properties, effects of surface reflectance, and

interference from clouds. Active sensing of aerosols with

lidar offers unique capabilities which are complementary to

those of passive techniques. Lidar is the only technique

giving high resolution profiles of aerosols, and it is able to

observe aerosol above bright surfaces, such as deserts and

snow, and above bright clouds. Because lidar provides its

Seasonal Variation of Aerosol Extinction

Coefficients and Aerosol Optical Depths Over

Bay of Bengal Region

G. V. Raghava Rayudu

Page 2: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2670

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

own illumination, aerosol can be observed over the full globe,

night as well as day yielding a more complete dataset for the

validation of regional and global aerosol models [3].

The data provided by the CALIPSO is processed to detect the

layers in the atmosphere whether it is aerosol or cloud and to

retrieve the extinction coefficient due to them. One drawback

of CALIPSO data is that the aerosol optical depth (AOD)

retrieved is not up to the precise mark because of the

uncertainties in the cloud -aerosol discrimination algorithm.

In order to find out the optical depth over that region, we go

for MODIS data which aboard on Aqua satellite.

Moderate resolution Imaging Spectroradiometer (MODIS)

is passive sensor aboard on Terra and Aqua satellites

launched by NASA. The MODIS instrument provides high

radiometric sensitivity in 36 spectral bands ranging in

wavelength from 0.4µm to 14.4µm. A 2330-km viewing

swath provides near-global coverage every day. Different

algorithms are used to retrieve AOD over ocean and over

land. Over ocean, seven wavelengths (0.47, 0.55, 0.66, 0.86,

1.2, 1.6, and 2.12µm) are used to retrieve aerosol optical

depth and other aerosol properties [4].

In this study, aerosol extinction profiles over Bay of Bengal

region are retrieved using CALIPSO satellite data for the

year 2012. Also, aerosol optical depths for the same region

are retrieved using MODIS-Aqua satellite data. Using the

retrieved data, the vertical and horizontal distribution of

aerosols over Bay of Bengal region are discussed for the four

seasons.

II. SATELLITE INSTRUMENTATION

CALIOP Instrument

The Cloud-Aerosol Lidar with Orthogonal Polarization

(CALIOP) is a two wavelength polarization lidar which

produces lidar pulse at 532 nm and 1064 nm. CALIOP laser

pulse makes a foot print of diameter 70 m on the earth with

the help of a beam expander. The laser pulse repetition

frequency is 20.16 Hz. Hence, it produces footprints every

335 m along the ground [2]. The instrument has a fixed

near-nadir view angle, so the measurements map a vertical

curtain along the orbital path and operates continuously,

acquiring 1.7 million laser shots every 24h and providing

observations during both day and night portions of the orbit.

The 0.532 m backscatter signal is sampled every 30m

vertically from 0.5 km to 8.2 km. Between 8.2 km and 20.2

km altitude profiles are averaged to 60m in the vertical and

every three successive shots are averaged together to give a

horizontal resolution of 1 km. The geolocated and

altitude-registered Level 1 data are calibrated before being

processed for Level 2 data products. Daytime measurements

have a lower signal-to-noise ratio than at night owing to the

noise added by the solar background illumination.

Figure II.1: CALIOP instrument [5]

The lidar sends two laser pulses, one at 532 nm and the other

1064 nm. If there is any layer of aerosol or cloud, we will

have backscattered pulse. This backscattered pulse is

detected by a telescope which is of 1 m in diameter. This

backscattered pulse is polarized light since we are sending a

laser beam, which is a polarized light [6]. Then by means of

a depolarizer, this backscattered pulse is depolarized and

thus we have light in multiple directions. After having the

depolarized light, an etalon is used to purify this light. Etalon

is device which narrows the bandwidth thereby we can get

the pulse at desired wavelength. The etalon is used only for

532 nm pulse is because of the low wavelength and it has a

large dynamic range. With the help of this etalon, the

interference takes place because of the depolarized light and

later the depolarized light is again polarized with the help of

a beam splitter. The purpose of the beam splitter is to separate

the polarized light into parallel and perpendicular

components.

MODIS Instrument

The Moderate Resolution Imaging Spectroradiometer

(MODIS) is a passive sensor which measures the radiances in

36 wavelengths from 0.41 to 14µm. The swath is 2330 km

and produces near global coverage every day [7]. The

radiation which is emitted from stars etc., rather than will be

below 4µm whereas the earth emits radiation greater than

4µm. This emitted radiation will be detected by the

instrument and thus parameters are retrieved. The parameter

we are concentrating is AOD. The channels which are used

to retrieve AOD over ocean 0.47, 0.55, 0.66, 0.86, 1.2, 1.6,

and 2.12µm have spatial resolutions of 250m or 500m a

calibration of the radiances is believed to be accurate to 2% or

better. Radiances are grouped into nominal 10-km cells

containing 20×20 pixels at 500 m resolution. Over land,

AOD is retrieved at 0.47, 0.55, and 0.66µm.

III. DATA COLLECTION AND RETRIEVAL

ALGORTIHMS

CALIPSO Data Collection

The data provided by the CALIOP instrument is processed

in three levels namely level-0, level-1 and level-2.

In general, Level 1 algorithms perform geolocation of the

lidar footprint and range determination, followed by

Page 3: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2671

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

determination of instrument calibration constants to produce

profiles of attenuated backscatter coefficients. The outputs of

the Level 1 algorithms are attenuated backscatter coefficient

profiles for the 532 nm and 1064nm intensity and for the

532nm perpendicular return, along with information on the

uncertainties in these products. These data are used by the

Level 2 algorithms to produce science data products. In

addition, calibration files are generated that track the

calibration constants that are determined and applied to the

data during Level 1 processing.

A. Level-0 Processing

All CALIPSO payload science data are downlinked once

per day to an X-band ground station in Alaska. The

X-band telemetry also includes payload health and status

data. The telemetry data are processed to level 0 data and

sent to the Atmospheric Sciences Data Centre (ASDC)

located at NASA Langley Research Centre in Hampton,

Virginia. Further processing, archiving, and distribution

of data from all three instruments are performed by

ASDC. Each day, approximately 3.5 GB of level 0

instrument science data are received from which 20 GB

of CALIOP level 1 and level 2 products are produced [6].

Automated analysis techniques must be able to identify

different optical properties of cloud and aerosol features

(e.g., the particulate extinction to backscatter ratio, also

commonly known as the lidar ratio) over different regions

of the globe as well as within the same column. The

processing system must be able to handle both features

easily detectable on a single-shot basis and also faint ones

that may require averaging a large number of profiles.

B. Level-1 Processing

Level 1 processing consists of three-dimensional Geolocation

followed by calibration.

a. Geolocation

The geometric location of each lidar footprint is

determined using spacecraft attitude and ephemeris data.

CALIPSO orbits in a circular orbit whereas the earth is an

oblate sphere, so the surface seems nearer at the poles and

farther at the equator [2]. As the satellite orbits the earth,

the range to the surface changes continually and the range

boundaries are automatically adjusted by the instrument.

During level-1 processing, the profiles are registered to a

common altitude grid using post processed satellite

ephemeris data.

b. Calibration

The level 1 products contain profiles of attenuated

backscatter, in units of km-1sr-1, but the received signal will

be in the units of power. Calibration consists of converting

the measured signal profile P(r), in digitizer counts, to

β’(r) 2 ( )

'( )A

r P rr

CEG (3.1)

Where C is the calibration coefficient, E is the measured

laser pulse energy, and GA is the electronic amplifier gain.

C. Level-2 Processing

The CALIOP Level 2 processing system is composed of

three modules, which are used for detecting layers,

classifying these layers by type, and performing extinction

retrievals [8]. These three modules are:

1. Selective Iterated Boundary Locator (SIBYL)

2. Scene Classifier Algorithm (SCA)

3. Hybrid Extinction Retrieval Algorithms (HERA)

Selective Iterated Boundary Locator (SIBYL):

The most fundamental data product from CALIOP is the

height of cloud and aerosol layers. SIBYL consists of an

algorithm to scan lidar profiles throughout the

troposphere and stratosphere, identifies regions of

enhanced scattering, and records the location and simple

characteristics of these atmospheric features and an

algorithm to average profiles and remove detected layers

from the profiles before further averaging [6]. Layers are

identified in the atmospheric return signal as

enhancements above the signal expected from the

molecular background. Identification is performed using

an adaptive thresholding technique applied to level one

532nm profiles β∥ + β⊥ , which generally provide better

sensitivity than the 1064nm channel. The level 1 profile

is converted into a profile of attenuated scattering ratio

R’(z)

'

'( )'( )

( )m

zR z

z

(3.2)

Before the thresholding algorithm is applied, the profile

of is estimated using gridded molecular and ozone

number density profile data from the Goddard Earth

Observing System Model, version 5 (GEOS-5) analysis

product available from the NASA Goddard GMAO [6].

Scene Classifier Algorithm (SCA):

The SCA is a set of algorithms that perform typing of the

detected layers based on layer height and layer

integrated properties [9]. In SCA, the first step is

determining whether a layer is cloud or aerosol using the

measured mean attenuated backscatter of that layer at

532 nm and the mean attenuated total color ratio of the

same, χ’ : '

1064

'

532

'

(3.3)

Where, the averages are computed from layer top to

base. Based on SCA, the layer is determined whether it

is an aerosol layer or cloud layer. SCA is composed of

three main sub modules [10].

i. Cloud-Aerosol Discrimination (CAD)

The CAD algorithm is a multidimensional

probability density function (PDF) based approach

[11].The PDFs used in the V2 data release are 3D,

including attributes (dimensions) of the mean

attenuated backscatter at 532 nm, the 1064/532

layer-integrated attenuated backscatter ratio (total color

ratio), and the midlayer altitude. This set of PDFs was

developed prelaunch based on existing airborne and

space borne lidar data sets. These PDFs which are in 3D

Page 4: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2672

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

space are given in a confidence function [12]

3 532( ' , ', )Df z as,

'

3 532( , ', )Df z

' '

532 532

' '

532 532

, ', , ',

, ', , ',

c a

c a

P z P z

P z P z

(3.4)

In the (3.5), cP and

aP are the PDFs for cloud and

aerosol, respectively. '

532 is the layer averaged

attenuated backscatter, χ’ is the total color ratio, z the

altitude. The function f is a normalized differential

probability that ranges from -1 to 1.

Figure III.1: Decision regions of aerosols and clouds [10]

There are a few scenarios which are frequently

misclassified in the V2 data release. A typical one is the

very dense dust layers associated with dust storms over

the source regions. Misclassifications occur because the

scattering properties of these dense dust layers are

nearly identical to those of optically thin clouds in the

3D space. So, in order to classify the dense dust layers

and optically thin clouds, we need to find one more

parameter which is known as volume depolarization

ratio (δV). The depolarization ratio is used to distinguish

between liquid and solid phases of water in the

atmosphere. In addition to δV, we find the latitude to

differentiate ocean with land. The increase in

dimension gets closer to the result [13]. Hence, the latest

version of CALIOP data i.e., in version 3, the CAD

algorithm is a set of PDFs in 5D, the dimensions are the

mean attenuated backscatter at 532 nm, the 1064/532

layer-integrated attenuated backscatter ratio (total color

ratio), and the midlayer altitude, the volume

depolarization ratio and the latitude. Therefore, the

confidence function in 5D space is given by [14],

'

5 532 , ', , ,D Vf z lat

' '

532 532

' '

532 532

, ', , , , ', , ,

, ', , , , ', , ,

c V a V

c V a V

P z lat P z lat

P z lat P z lat

(3.5)

Where, δV is the volume depolarization ratio and lat is

the latitude.

The CAD score reported in the CALIPSO L2 products is

a percentile (integer) of f ranging from -100 to 100. A

feature is classified as cloud when 0f and as aerosol

when 0f . The absolute value of the CAD score

provides a confidence level for the classification [15].

ii. Aerosol subtyping

After clouds are separated from aerosol layers,

SCA attempts to identify the type of aerosol in the layer.

By type, we mean a mixture of aerosol components that

is characteristic of a region. An aerosol lidar ratio,

where σa is the aerosol extinction coefficient and βa is

aerosol back scatter coefficient, is associated with each

aerosol type so that identification of aerosol type

provides a value of Sa to be used in the extinction

retrieval [6]. Aerosol type is also useful in its own right,

because type identification is a step in the process of

identification of the aerosol source and attribution of

aerosol radiative forcing to natural or anthropogenic

emissions. Proper lidar ratio is required to be selected in

order to correctly classify the aerosol type. More details

about lidar ratio can be seen in the next section where

we discuss about the lidar equation.

iii. Cloud Ice-Water Phase Discrimination

The CALIOP laser transmits a pulse that is nearly

100 percent linearly polarized. The ratio of the

perpendicular and parallel backscatter returns is the

volume depolarization [16], '

'V

Which determines whether the laser pulse has been

scattered by liquid cloud droplets or by ice crystals. This

algorithm is required in discriminating the ice particles

from the clouds. More details about Cloud Ice-Water

Phase Discrimination will be available from [16].

Hybrid Extinction Retrieval Algorithms (HERA)

HERA retrieves profiles of particle backscatter and

extinction by using a simple iterative algorithm which is

more convenient to retrieve extinction coefficient [17].

The retrieval assumes a two-component atmosphere i.e.,

molecules and particles. The particles are nothing but

aerosols. Solving for the unknown particulate extinction

and backscatter requires either an a priori lidar ratio or

the layer transmittance as a boundary condition. To

generate the level 2 profile products, the retrieved

profiles are averaged or replicated, as necessary, to a

common resolution of 40km horizontal and 120m

vertical for the aerosol profile product and replicated to

5km horizontal and 60m vertical for the cloud profile

product [8].

HERA is designed to perform extinction retrievals by

using results from the SIBYL and SCA algorithms, and

it is designed with considerable flexibility because a

variety of different scenarios must be considered. When

the transmittance of a layer can be derived from the

clear-air signal above and below the layer, it is used as a

boundary condition on the retrieval. These constrained

retrievals tend to be the most accurate retrievals, and

this method also retrieves a representative value of the

lidar ratio for the layer. When a transmittance

measurement is not possible (e.g., boundary layer

Page 5: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2673

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

aerosol or opaque cloud), an a priori lidar ratio must be

used. For aerosols, the lidar ratio comes from the SCA

module. The extinction coefficient is retrieved using the

solution of the lidar equation.

Solution of lidar equation:

The lidar equation is given by [18],

2 2

2( ) ( ) ( ) ( ) ( )m a m a

EcP z z z T z T z

z (3.6)

0

2 ( )2 ( )

z

m z dz

mT z e

(3.7)

0

2 ( )2 ( )

z

a z dz

aT z e

(3.8)

Where E is the energy of the pulse sent, c is the

calibration constant which depends upon factors such as

receiver cross section; efficiency of the detector and

optical system; and corrections for near-range

field-of-view problems, βm is the molecular back scatter

coefficient and βa is the aerosol back scatter coefficient,

Tm2 is the molecular two-way transmittance and Ta

2 is the

aerosol two-way transmittance [19]. The molecular

back scatter coefficient and the molecular two-way

transmittance are known because of the availability of

the size distribution of molecules, but in case of

aerosols, it varies. Hence the above equation becomes an

equation of two unknowns i.e., the aerosol back scatter

coefficient and aerosol extinction coefficient. In order to

solve this equation, we assume a quantity known as lidar

ratio which is the ratio of extinction coefficient and back

scatter coefficient at 532 nm.

( )

( )

zS

z

(3.9)

The molecular lidar ratio is known, since βm the σm and

are known, but the βa and σa are not known, as σa is the

quantity we need to retrieve, so we need to find aerosol

lidar ratio. The molecular lidar ratio is 8

3mS

and

the aerosol lidar ratio is given as,

( )

( )

a

a

zS

z

From (3.7), let ( ) ( ) ( )m az z z and

2 2 2( ) ( ) ( )m aT z T z T z

0

2 ( )2 ( )

z

z dz

T z e

( )

( ) ( )( )

zS z S z

z

0

2 ( )2 ( )

z

S z dz

T z e

(3.10)

Equation (3.6) is solved using Klett’s inversion method

[20].

Now apply logarithm on both sides,

2

0

ln ( ) 2 ( )z

T z S z dz (3.11)

2

2

1 ( )2 ( )

( )

dT zS z

T z dz (3.12)

2

2

1 ( )( )

2 ( )

dT zz

ST z dz (3.13)

Since from (3.6),

2

2( ) ( ) ( )

EcP z z T z

z

Therefore from (3.13),

22

2 2

1 ( )( ) ( )

2 ( )

Ec dT zP z T z

z ST z dz

2

2

( )( )

2

Ec dT zP z

Sz dz

Now, solve for T2(z), 2

2

0

2( ) ( )

zSz

dT z P z dzEc

(3.15)

We assume the integration constant to be 1. Therefore,

2 2

0

2( ) 1 ( )

zS

T z z P z dzEc

(3.16)

22

0

2 ( )( ) 1

zS z P z

T z dzc E

(3.17)

From (3.7),

2

2( ) ( ) ( )

EcP z z T z

z (3.18)

21

2( )( ) ( )

z P zz T z

Ec

(3.19)

2 2

0

( ) 2 ( )( ) 1

zz P z S z P z

z dzEc c E

(3.20)

22

0

2 ( )( ) 1

zS z P z

T zc E

(3.21)

* 22

0

2 ( )1 ( *)

zS z P z

dz T zc E

(3.22)

z* is a reference calibration range i.e., where

( ) ( *)z z

* 22

0

( )1 ( *) 2

zz P z

S c T z dzE

(3.23)

From (3.7),

22 2

2 2

( )1( ) ( ) ( ) ( )

2 ( )

am m a

a

dT zEcP z z T z T z

z ST z dz

22

2 2

2 2

( )2 ( ) ( )

( ) ( ) ( )2 ( )

aa m

m a

a

dT zST z z

dzEcP z T z T z

z ST z

(3.24)

Page 6: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2674

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

22 2 2 2

2

( )2 ( ) ( ) 2 ( ) ( ) ( ) ( )a

a a m m a

dT zEcST z P z ST z z T z T z

z dz

(3.25)

222

2

( )2 ( )2 ( ) ( )

( )

aa

m

dT zSz P zST z z

EcT z dz (3.26)

2 22

2

( ) 2 ( )2 ( ) ( )

( )

aa m

m

dT z Sz P zST z z

dz EcT z (3.27)

(3.27) is of the form ' ( ) ( )y a x y b x , which has the

solution,

( )

1

( )

( )a x dx

a x dx

e b x dx Cy

e

(3.28)

Here,2 ( )ay T z , ( ) 2 ( )ma x S z ,

2

2

2 ( )( )

( )m

Sz P zb x

EcT z and

assume1 1C .

Therefore, 2

2 ( )

2

2

2 ( )

2 ( )1

( )( )

m

m

S z dz

m

a S z dz

Sz P ze

EcT zT z

e

(3.29)

22 ( ) 2 ( )2

2

2 ( )( ) 1

( )

m mS z dz S z dz

a

m

S z P zT z e e dz

c ET z

(3.30)

Thus 2 ( )aT z is determined from (3.30). And from (3.9),

( )a z can be retrieved. After retrieving the aerosol

extinction coefficient, the column integrated aerosol optical

depth is given by [21].

0

(0, ) ( )z

z z dz (3.31)

MODIS Data Collection:

The data processed in MODIS follows two algorithms. One

for land and the other for ocean. Aerosols over also are

difficult to observe because of etc., whereas on water, it is

easier to identify. So the MODIS use complex algorithm for

observations over land. Further details in MODIS retrieval

algorithms can be referred from [22] [23].

IV. DATA ACQUISITION

The data [24] considered is over Bay of Bengal region

between the latitudes 0oN to 10oN and longitudes 80oE to

95oE in the year 2012. The grid selected covers southern part

of Bay of Bengal. The main reason to consider this part is that

this is the region where depression occurs and cyclones enter

Indian subcontinent [25].

The grid is divided into 15 sub-grids such that for every 2o

latitude, we divide longitude from 80o to 95o in three equal

parts. The profiles are then retrieved by averaging for every

day and then averaging all the days in month and then

averaging by seasons.

Figure IV.1: Region of interest

The above grid is divided into 15 sub-grids in order to

investigate over small interval of region as the aerosol

concentration vary rapidly for several hundred miles, the

study concentrates in the seasonal variation of aerosol

extinction profiles that are retrieved from the data of

CALIPSO and also the aerosol optical depths (AOD) using

MODIS data. The averaged profiles are observed in the four

seasons namely winter (December-February), spring

(March-May), summer (June-September) and autumn

(October-November).

One can raise a question that, since we retrieval aerosol

extinction coefficient using CALIPSO data, why don’t we

consider the aerosol optical depths from the same data itself?

The reasons are as follows [26]:

(1) The extinction profiles that are retrieved from

CALIPSO data are limited to a certain altitude.

(2) The uncertainties in cloud-aerosol discrimination

algorithm.

Because of these two reasons, we do not consider the

AOD’s from CALIPSO data. One alternative is to consider

AOD data from MODIS instrument aboard on Aqua

satellite. The AOD’s retrieved from MODIS, uses

different algorithms over land and ocean. Aqua and

CALIPSO belong to the same group of satellites A-train

and have almost same equatorial crossing times, so data

between the two satellites are collocated.

V. RESULTS

The altitude range fixed for processing the data is from 0 km

to 4km. This chosen so, because the AOD calculated using

CALIPSO’s aerosol extinction values shows nearly values as

provided in the look up table of the hdf file in which the

aerosol extinction is actually columnar integrated from 0 km

to 30 km.

The AOD which has been retrieved by CALIPSO is quite

different from that of MODIS retrieved AOD. This is because

of the uncertainties induced during the processing of

CALIPSO level-2 data. The uncertainties are due to the

classification of aerosols and clouds in CAD algorithm and

the assumption of aerosol lidar ratio in retrieval of aerosol

extinction coefficient. Hence, MODIS data is chosen for

AOD.

Page 7: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2675

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

Figure V.1: Scatter plot of AOD between MODIS and CALIPSO

Fig. V.1 shows the scatter plot between AODs from

MODIS and CALIPSO. The plot the AODs are not the

same in the two sources. The points are monthly averaged

AODs from MODIS and CALIPSO.

The seasonal variation of mean aerosol extinction

coefficient is given below. The profiles as shown per fig.

IV. 1. All profiles are averaged for every season.

Mean Extinction Profiles in winter:

Figure V.2: Mean extinction profiles in 8

oN-10

oN and 80

oE-95

oE in winter

Figure V.3: Mean extinction profiles in 6

oN-8

oN and 80

oE-95

oE in winter

Page 8: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2676

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

Figure V.4: Mean extinction profiles in 4

oN-6

oN and 80

oE-95

oE in winter

Figure V.5: Mean extinction profiles in 2

oN-4

oN and 80

oE-95

oE in winter

Figure V.6: Mean extinction profiles in 0

oN-2

oN and 80

oE-95

oE in winter

Mean Extinction Profiles in spring:

Figure V.7: Mean extinction profiles in 8

oN-10

oN and 80

oE-95

oE in spring

Figure V.8: Mean extinction profiles in 6

oN-8

oN and 80

oE-95

oE in spring

Page 9: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2677

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

Figure V.9: Mean extinction profiles in 4

oN-6

oN and 80

oE-95

oE in spring

Figure V.10: Mean extinction profiles in 2

oN-4

oN and 80

oE-95

oE in spring

Figure V.11: Mean extinction profiles in 0

oN-2

oN and 80

oE-95

oE in spring

Mean Extinction Profiles in summer:

Figure V.12: Mean extinction profiles in 8

oN-10

oN and 80

oE-95

oE in summer

Figure V.13: Mean extinction profiles in 6

oN-8

oN and 80

oE-95

oE in summer

Page 10: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2678

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

Figure V.14: Mean extinction profiles in 4

oN-6

oN and 80

oE-95

oE in summer

Figure V.15: Mean extinction profiles in 2

oN-4

oN and 80

oE-95

oE in summer

Figure V.16: Mean extinction profiles in 0

oN-2

oN and 80

oE-95

oE in summer

Mean Extinction Profiles in autumn:

Figure V.17: Mean extinction profiles in 8

oN-10

oN and 80

oE-95

oE in autumn

Figure V.18: Mean extinction profiles in 6

oN-8

oN and 80

oE-95

oE in autumn

Page 11: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2679

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

Figure V.19: Mean extinction profiles in 4

oN-6

oN and 80

oE-95

oE in autumn

Figure V.20: Mean extinction profiles in 2

oN-4

oN and 80

oE-95

oE in autumn

Figure V.21: Mean extinction profiles in 0

oN-2

oN and 80

oE-95

oE in autumn

Variation of mean AOD in winter, spring, summer, autumn:

Latitude Longitude Seasons

Winter Spring Summer Autumn

0oN-2oN

80oE-85oE 0.0587 0.0712 0.1044 0.0462

85oE-90oE 0.0544 0.0590 0.0991 0.0458

90oE-95oE 0.0567 0.0536 0.0723 0.0453

2oN-4oN

80oE-85oE 0.0691 0.0829 0.1420 0.0386

85oE-90oE 0.0551 0.0706 0.1064 0.0440

90oE-95oE 0.0494 0.0545 0.0819 0.0385

4oN-6oN

80oE-85oE 0.0874 0.1151 0.1642 0.0478

85oE-90oE 0.0744 0.0998 0.1435 0.0509

90oE-95oE 0.0626 0.0780 0.1040 0.0467

6oN-8oN

80oE-85oE 0.0977 0.1283 0.1845 0.0566

85oE-90oE 0.0846 0.1216 0.1906 0.0576

90oE-95oE 0.0818 0.1048 0.1395 0.0588

8oN-10oN

80oE-85oE 0.0879 0.1306 0.1647 0.0614

85oE-90oE 0.0891 0.1294 0.2213 0.0660

90oE-95oE 0.0804 0.1281 0.1727 0.0542

Mean AOD 0.0726 0.0952 0.1394 0.0506 Table V.1: Variation of mean AOD in winter, spring, summer and autumn

From the profiles, we can observe that, in summer and

autumn, thick layers of aerosols are formed. Thick layers

of aerosols formed in summer could be due to evaporation

of sea water, particles such as sea salt can be lifted along in

the evaporation process. The aerosol layers in autumn

could show the presence of clouds. Maximum extinction

Page 12: Seasonal Variation of Aerosol Extinction Coefficients and ...ijsetr.org/wp-content/uploads/2015/07/IJSETR-VOL-4... · Indian Institute of Space Science and Technology, Hyderabad,

International Journal of Science, Engineering and Technology Research (IJSETR), Volume 4, Issue 7, July 2015

2680

ISSN: 2278 – 7798 All Rights Reserved © 2015 IJSETR

often occurred between 0oN to 4oN latitudes. The altitude

is fixed to 4km, because the column integrated aerosol

optical depths calculated from CALIPSO data shown same

values as of the aerosol optical depths calculated from 0 to

4km.

The monthly averaged aerosol extinction profiles shows

how the AOD and aerosol extinction is varying seasonally.

In winter, the average AOD is 0.0726. Where as in spring,

the AOD is increased to 0.0952. The maximum AOD is

0.1394 which is in summer and the minimum is in autumn

which AOD happened to be 0.0506.

VI. CONCLUSIONS

The CALIPSO and MODIS data has been downloaded

and studied briefly about the Science Data Sets provided in

the data. The CALIPSO data is processed using matlab to

produce the monthly averaged extinction profiles in the

year 2012 and the monthly averaged aerosol optical depth

values retrieved from MODIS data have been included in

the profiles using the same.

The aerosol optical depth that has been retrieved during

the processing CALIPSO Level 2 data and the AOD values

calculated from extinction profiles shows that the aerosol

layers are concentrated below 4 km altitude over the Bay of

Bengal region and studied about the seasonal variation of

aerosol extinction and aerosol optical depth.

The aerosol extinction profiles shows aerosol layers are

most commonly concentrated below 2 km altitude i.e.,

within the Atmospheric Boundary Layer (ABL) and the

profiles which are showing above 2 km says that the

aerosol layers has been lifted up due to circulation. More

aerosol extinction has been observed in summer and

autumn and more AOD has been observed in summer.

ACKNOWLEDGMENT

I would like to express my sincere and deep gratitude to

my guide, Dr. M. V. Ramana, Associate Professor,

Indian Institute of Space Science and Technology, who

constantly supported my research work throughout the

project. I would also like to thank Mr. Himanshu Singh, a

colleague of mine, who motivated me to complete this

project. Finally, I would like to thank my family, who

stood beside me in hard times.

REFERENCES

[1] Wallace, John M and Hobbs, Peter V. Atmospheric science: an

introductory survey, Vol. 92(169), 2006.

[2] Winker D M. Overview of the CALIPSO Mission and CALIOP Data

Processing Algorithms, J. Atmos. Oceanic Technol., VOL. 26,

2310-2320, 2009.

[3] Winker D M. The CALIPSO mission: space borne lidar for

observation of aerosols and clouds, Remote Sens. Atmos. Ocean,

Environment Space, 2003.

[4] Kittaka C. Intercomparision of column aerosol depths from CALIPSO

and MODIS-Aqua, J. Atmos. Meas. Tech., VOL. 4, 131-134, 2011

[5] Winker D M. Status and performance of the CALIOP lidar, J. Remote

Sens., 3, 2004.

[6] Winker, D, M. CALIOP Algorithm Theoretical Basis Document Part

1: CALIOP Instrument, and Algorithms Overview, 2006.

[7] Salomonson, V. Vincent. An Overview of the Earth Observing System

MODIS Instrument and Associated Data Systems Performance, IEEE,

1174-1175, 2002.

[8] Young, A. Stuart. The Retrieval of Profiles of Particulate Extinction

from Cloud -Aerosol Lidar Infrared Pathfinder Satellite Observations

(CALIPSO) Data: Algorithm Description. J. Atmos. Oceanic

Technol. 1105-1119, 2008.

[9] Liu, Z. The CALIPSO Lidar Cloud and Aerosol Discrimination:

Version 2 Algorithm and Initial Assessment of Performance, J. Atmos.

Oceanic Technol., VOL. 26, 1198-1207, 2009.

[10] Liu, Z. The calipso cloud and aerosol discrimination: version 3

algorithm and test results, J. Atmos. Oceanic Technol., Vol. 26,

1918-1213, 2009.

[11] Liu, Z. Use of probability distribution functions for discriminating

between cloud and aerosol in lidar backscatter data, J. Geophys. Res.,

Vol. 109, 2004.

[12] Vaughan, M. fully automated analysis of space-based lidar data: an

overview of the CALIPSO retrieval algorithms and data products, J.

Appl. Meteorol., Vol. 5575, 2004.

[13] Liu, Z. A Cloud-Aerosol Discrimination Algorithm for CALIPSO

Lidar Observation: Algorithm Tests, ILRC 2004

[14] Hu, Yongxiang. CALIPSO/CALIOP cloud phase discrimination

algorithm, J. Atmos. Oceanic Technol. Vol.26, 2293-2309, 2009.

[15] Weidong, Y. Effect of CALIPSO cloud-aerosol discrimination (CAD)

confidence levels on observations of aerosol properties near clouds, J.

Atmos. Res., VOL. 116, 134-141, 2012.

[16] Yongxiang Hu, CALIPSO/CALIOP Cloud Phase Discrimination

Algorithm, J. Atmos. Oceanic Technol., 2009.

[17] Vaughan, Mark A. Adapting CALIPSO Climate Measurements for

Near Real Time Analyses and Forecasting, 2011.

[18] Fernald, G. Frederick. Determination of Aerosol Height Distributions

by Lidar, J. APPL. METEOROL., VOL. 11, 483-484, 1971.

[19] Stuart A. Young. CALIOP Algorithm Theoretical Basis Document

Part 4: Extinction Retrieval Algorithms. Tech. Rep. 2008.

[20] Walter Carnuth. Cloud extinction profile measurements by lidar using

Klett’s inversion method, J. APPL. OPT., VOL. 25, 2899-2901, 1971.

[21] Chan, P. W. Determination of Backscatter-Extinction Coefficient

Ratio for LIDAR-Retrieved Aerosol Optical Depth Based on

Sunphotometer Data, J. Remote Sens., VOL. 2, 2129, 2010.

[22] Remer, L, A. The MODIS aerosol algorithm, products, and validation.

J. atmos. sci., Vol. 62, 947-973. 2005.

[23] Wang, J. Improved algorithm for MODIS satellite retrievals of aerosol

optical thickness over land in dusty atmosphere: Implications for air

quality monitoring in China, J. Remote Sens. Environment, Vol. 114,

2575-2583, 2010.

[24] Website to download CALIPSO and MODIS data (accessed

February, 2015): http://reverb.echo.nasa.gov

[25] Daniel, R. Aerosol Effects on Microstructure and Intensity of Tropical

Cyclones, Bull. Amer. Meteor. Soc., Vol. 93, 2012.

[26] Xiaoyan Ma. Comparison of AOD between CALIPSO and MODIS:

significant differences over major dust and biomass burning regions, J.

Atmos. Meas. Tech., VOL. 6, 2391-2392, 2013.

[27] Zhang. X. Q., RETRIEVAL OF AEROSOL OPTICAL DEPTH

OVER URBAN AREAS USING TERRA/MODIS DATA, J. Remote

Sens., VOL. 38, 374-375, 2010.

[28] Humera, B. Intercomparision of MODIS, MISR, OMI, and CALIPSO

aerosol optical depth retrievals for four locations on the Indo-Gangetic

plains and validation against AERONET data. J. Atmos.

Environment, Vol. 111, 113-126, 2015.

Author Profile

G. V. Raghava Rayudu completed B.Tech in physical sciences, from

Indian Institute of Space Science and Technology, has done project in

atmospheric sciences entitled ―Retrieval of Aerosol Extinction Profiles and

Aerosol Optical Depths over Bay of Bengal Region‖. This paper is based on

the project mentioned above.