Investigation on the spatial distribution of PM 2 5 by integraring satellite image from 2013 2015

LIST OF ABBREVIATIONS planetary boundary layer Aerosol Optical Depth Total column AOD AOD under 1000 m AOD under 500 m AOD under 200 m APD under 140 m AOD under 70 m Particle mass with d

Trang 1

UNIVERSITY OF AGRICULTURE AND FORESTRY

HA THI HONG INVESTIGATION ON THE SPATIAL DISTRIBUTION OF PM 2.5 BY INTEGRATING SATELLITE IMAGE FROM 2013-2015

BACHELOR THESIS

Study Mode : Full-time

Major : Environmental Science And Management Faculty : International Training and Development Center Batch : 2012 - 2016

Thai Nguyen, December 2016

Trang 2

Thai Nguyen University of Agriculture and Forestry

Degree Program Bachelor of Environmental Science and Management Student name HA THI HONG

Previous studies reported that human health is strongly and consistently affected by outdoor fine mode particulate matter, the so-called PM2.5 The monitor of PM concentration thus attracts much attention for the society However, the observations for a spatial distribution are limited to the location of ground based stations Therefore, Satellite images with a wide coverage like MODIS, MISR and OMI have been applied to overcome this limitation in terms of retrieved AODs

For the application of total column AOD to PM concentration, the vertical distribution and aerosol type should be taken into account Association with extinction profile of CALIPSO products and aerosol compositions from NGAI algorithm, the AOD retrievals and PM2.5 are correlated in this study As the result, the linear regression between AOD from CALIPSO and PM2.5 has more uncertainty if AODtotal, AOD1000 and AOD500 used as proxy to estimate PM2.5 concentration (R2<0.2) compared to lower altitudes AOD200, AOD140 and AOD70 The result concluded that AOD140 from CALIPSO can be used as a better proxy for PM2.5 concentration (R2

Trang 3

range from 0.58 to 0.79) Using NGAI algorithm classified aerosol compositions for AERONET and validated to CALIPSO subtypes indicated that dominant aerosol type (polluted continent) in study area is much improved AOD140 - PM2.5 relationship (R2range between 0.86 and 0.99)

Keywords PM2.5, AOD, CALIPSO, Aerosol types, NGAI, Aerosol

layers Number of pages: 45

Date of submission: 03/12/2016

Trang 4

From bottom of my heart, I would like to express my deepest appreciation to Associate Professor Tang-Huang Lin who in spite of being extraordinarily busy with his duties, took time out to hear, guide, keep me on the correct path and complete report during the time of conducting the research at Center for Space and Remote Sensing Research (CSRSR) of National Central University (NCU)

I also wish to express my deep gratitude to MSc Nguyen Van Hieu who gives

me an opportunity, guidance and support me to complete thesis I would also like to express my great appreciation to Mr Wei Hung Lien and Ms Chang Yi-Ling for their constant support, patient guidance and suggestions related to my work I sincerely thank the additional members of Center for Space and Remote Sensing Research who have contributed to my work Last but not the least, I would like to thank all of my family members and dear friends who always encourage and back me up unceasingly

Thai Nguyen, December 2016

HA THI HONG

Trang 5

TABLE OF CONTENTS

LIST OF TABLES xi

LIST OF ABBREVIATIONS xii

PART I.INTRODUCTION 1

1.1 Research rationale 1

1.2 Research’s questions 3

1.3 The requirement 3

PART II LITERATURE REVIEW 4

2.1 Ground based Measurements 4

2.1.1 Ground based Measurement - PM2.5 4

2.1.2 Ground based measurement - AERONET 6

2.2 Brief Description of Remote Sensing – CALIPSO 7

2.3 The Aerosol particles and Normalized Gradient Aerosol Index (NGAI) 11

2.3.1 The role of aerosol types 11

2.3.2 Normalized Gradient Aerosol Index (NGAI) 12

2.3.3 AOD fraction of mixed type aerosols 14

2.3.4 Study area 16

2.3.5 Software 16

PART III DATA AND METHODOLOGY 17

3.1 Data 17

3.1.1 AERONET - Ground based measurement 17

3.1.2 PM2.5 stations 19

Trang 6

3.1.3 Satellite data 20

3.2 Methodology 22

PART IV RESULT AND DISCUSSION 26

4.1 Correlations between hourly PM2.5 and AOD in various layers 26

4.3 Aerosol types classification and AOD Fraction Determination 33

PARTV CONCLUSION AND SUGGESSION 35

REFERENCES 37

Trang 7

LIST OF FIGURES

Figure 1: Size comparison between two aerosols with diameters 2.5 and 10 µ m, a

human hair and a sand grain (credit: Environmental Protection Agency) 12

Figure 2: The orbit track of CALIPSO passes to Taiwan at 17:40pm on August 10,

2014 16

Figure 3: Total Attenuated Backscattering signal measured by the CALIOP passed

Taiwan (red box) level 2 at 532 nm during the period 17:40- 17:54 UTC p 17

Figure 4: Vertical Feature Mask measured by the CALIOP passed Taiwan (red box)

level 2 during the period 17:40- 17:54 UTC 18

Figure 5: Aerosol subtype classification information measured by the CALIOP passed

Taiwan (red box) level 2 during the period 17:40- 17:54 UTC 20

Figure 6: Aerosol extinction coefficient profile from CALIPSO at 12:54:20 (LZT) on

March 27, 2013 in Taipei city 21

Figure 7: The scheme of AOD fraction determination for dual-type aerosols (type A

and B) based on NGAI values (Lin et al., 2016) 25

Figure 8: The locations of AERONET site and PM2.5 stations selected in this study

(google map) 28

Figure 9: The flowchart of analysis procedure 30

Figure 10a: The linear regression between AOD140 – PM2.5 in Guting from 2013 to

Formatted: Font: 14 pt, Bold, Font color: Auto

Formatted: Font color: Auto

Formatted: Justified, Line spacing: Double

Trang 8

Figure 11b: The improved correlation between AOD140 and PM2.5 in Tucheng from

Figure 12: The NGAI identification result without AOD fraction in Taipei_WCB 39

Figure 13: The NGAI identification result with AOD fraction in Taipei_WCB 40

Figure 1: Size comparison between two aerosols with diameters 2.5 and 10 µ m, a

human hair and a sand grain (credit: Environmental Protection Agency) 15

Figure 2: The orbit track of CALIPSO passes to Taiwan at 17:40pm on August 10,

2014 19

Figure 3: Total Attenuated Backscattering signal measured by the CALIOP passed

Taiwan (red box) level 2 at 532 nm during the period 17:40- 17:54 UTC p 20

Figure 4: Vertical Feature Mask measured by the CALIOP passed Taiwan (red box)

level 2 during the period 17:40- 17:54 UTC 21

Figure 5: Aerosol subtype classification information measured by the CALIOP passed

Taiwan (red box) level 2 during the period 17:40- 17:54 UTC 23

Figure 6: Aerosol extinction coefficient profile from CALIPSO at 12:54:20 (LZT) on

March 27, 2013 in Taipei city 24

Figure 7: The scheme of AOD fraction determination for dual-type aerosols (type A

Formatted: Justified

Trang 9

Figure 8: The locations of AERONET site and PM2.5 stations selected in this study

(google map) 31

Figure 9: The flowchart of analysis procedure 33

Figure 10a: The linear regression between AOD140 – PM2.5 in Guting from 2013 to

Figure 13: The NGAI identification result with AOD fraction in Taipei_WCB 42

Figure 1: Size comparison between two aerosols with diameters 2.5 and 10 µ m, a human hair and a sand grain (credit: Environmental Protection Agency) 5Figure 2: The orbit track of CALIPSO passed to Taiwan at 17:40pm on August 10,

2014 (Source: www-calipso.larc.nasa.gov) 9

Trang 10

Figure 3: Total Attenuated Backscattering signal measured by the CALIOP passed Taiwan (red box) level 2 at 532 nm during the period 17:40- 17:54 UTC (Source: www-calipso.larc.nasa.gov) 10Figure 4: Vertical Feature Mask measured by the CALIOP passed Taiwan (red box) level 2 during the period 17:40- 17:54 UTC (Source: www-calipso.larc.nasa.gov) 11Figure 5: The scheme of AOD fraction determination for dual-type aerosols (type A and B) based on NGAI values (Lin et al., 2016) 15Figure 7: The location of AERONET site and PM2.5 stations selected in this study (Google map) 19Figure 8: Aerosol subtype classification information measured by the CALIOP passed Taiwan (red box) level 2 during the period 17:40- 17:54 UTC 21Figure 9: Aerosol extinction coefficient profile from CALIPSO at 12:54:20 (LZT) on March 27, 2013 in Taipei city 22Figure 10: The flowchart of analysis procedure 25Figure 11a: The linear regression between AOD140 – PM2.5 in Guting from 2013 to

2015 28Figure 11b: The linear regression between AOD140 – PM2.5 in Tucheng from 2013 to

2015 28Figure 11c: The linear regression between AOD140 – PM2.5 in Zhonghe from 2013 to 2015……… 31 Figure 11d: The linear regression between AOD140 – PM2.5 in Xindian from 2013 to 2015……… 32 Figure 11e: The linear regression between AOD140 – PM2.5 in Banqiao from 2013 to

Trang 11

Figure 13: The NGAI identification result without AOD fraction in Taipei_WCB 33

Figure 14: The NGAI identification result with AOD fraction in Taipei_WCB 37 Formatted: Hyperlink, Font: Not Italic, Font

color: Auto, Do not check spelling or grammar

Formatted: Hyperlink, Font: Not Italic, Font color: Auto, Do not check spelling or grammar

Formatted: Justified, Space After: 0 pt

Trang 12

LIST OF TABLES

Table 1: PM2.5 concentration and air pollution banding 6Table 2: Aerosol classification using NGAI algorithm 14Table 3: The NGAI threshold of AP, BB and Dust aerosols used to classify aerosol types in AERONET site 19Table 4: CALIPSO data collected in this study 20Table 5: Correlations between hourly PM2.5 and AOD from 500m to total column AOD 26Table 6: Correlations between hourly PM2.5 and AOD from ground surface to 200m 27Table 7: The improved correlation between AOD and PM2.5 30

Trang 13

LIST OF ABBREVIATIONS

planetary boundary layer

Aerosol Optical Depth Total column AOD AOD under 1000 m AOD under 500 m AOD under 200 m APD under 140 m AOD under 70 m Particle mass with diameters less than 2.5 mm

The Normalized Gradient Aerosol Index

Light Detection and Ranging Cloud-Aerosol Lidar and Infrared Pathfinder Satellite

Observation

AODtotal Total column AOD

Formatted: Font: 14 pt, Bold, Font color: Auto

Formatted: Normal, Line spacing: Double

Formatted Table Formatted: Font color: Auto

Formatted: Font color: Auto, Subscript

Trang 14

AOD500 AOD under 500 m

CALIPSO Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation

Lidar Light Detection and Ranging

NGAI Normalized Gradient Aerosol Index

PM2.5 Particle mass with diameters less than 2.5 mm

Trang 15

PART I

INTRODUCTION 1.1 Research rationale

Air pollution exerts significant impacts influences on environments, visibility

and human health When particle matterparticulate matter (PM) concentrations become

high enough, they can pose serious health risks, especially to individuals with asthma

and other respiratory problems as well as affect transmission of solar radiation through

scattering and absorption (Nwafor et al., 2007) Airborne aerosols can also transport

fungal and viral microbial pathogens , which can lead to disease outbreaks in other

parts of the world

Previous studiesy concluded pointed out that human health is strongly and

consistently affected by outdoor fine particle matterparticulate matter than coarse

particle matterparticulate matter (PM10), an increase of 50 µm/m3 in the concentration

causes 1–8% more increase of deaths (Wallace, 2000) as well as cause serious

respiratory and cardiovascular diseases that lead to the premature mortality (Dockery,

D.W., & Pope, 1994; Krewski et al., 2000; Pope et al., 2002; Künzli et al., 2005;

Brook et al., 2010) PM2.5 is considered as an important index of air pollutions;

especially it is one of the major air pollutants observed in the past decade in Taiwan

(Taiwan Environmental Protection Administration, 2008) PM2.5 is also well known

as smaller particles with aerodynamic diameters less than 2.5 mm., T the total mass

concentration of fine particles and is measured in ground-based measurement

Although theose ground-based measurements are relatively accurate, they are

representative of a limited area because aerosol sources could vary over small spatial

scales and the aerosol lifetime is less than an hour to a few days, depending on particle

Formatted: Section start: New page, Width:

21 cm, Height: 29.7 cm, Top: (No border), Bottom: (No border), Left: (No border), Right: (No border), Not Different first page header

Formatted: Font color: Auto, Pattern: Clear

Formatted: Font: 14 pt, Bold

Formatted: Font: Bold, Font color: Auto

Formatted: Indent: First line: 1.27 cm

Formatted: Font color: Auto, Superscript

Trang 16

size and chemical compositions (Schaap et al., 2008) Moreover, the large spatial and

temporal variability of airborne particles makes difficult to estimate the abundance at

any given locations based upon limited surface observations (Kumar et al., 2011)

Several studies have focused on correlating satellite AOD observations and PM2.5

concentrations by the most widely used total column AOD satellite aerosol products,

such as the Dark Target (DT)/Deep Blue (DB) MODIS and Multi-angle Imaging

Spectroradiometer (MISR; Diner et al., 1998; Kahn et al., 2010) aerosol products (e.g.,

Shi et al., 2011b) in order to overcome such limitations and provide information of aerosol

particles in the lower troposphere near the surface and monitoring aerosol concentration at

global/ regional scale

However, that needs be considered when applying satellite-based observations

in general, much less as a proxy for PM2.5 estimates First, uncertainties exist in

satellite retrieved AOD values due to issues such as cloud contamination, inaccurate

optical models used in the retrieval process , and heterogeneous surface boundary

conditions (Toth et al., 2014) Any estimate of PM2.5 derived from satellite AOD data

cannot be more accurate than the AOD data themselves Thus, relationships between

AOD and PM2.5 are likely to be highly sensor specific production Second, AOD

derived from passive sensors is a column integrated value, and PM2.5 concentration is

a surface measurement Under conditions where aerosol particles are concentrated

primarily within the surface/boundary layer, AOD is presumably a likelier proxy for

PM2.5 concentration Finally, AOD is a column-integrated sum of total ambient

particle extinction, whereas PM2.5 is measured with respect to dried particles ingested

for analysis by corresponding instruments

Formatted: Font color: Auto, Condensed by 0.2 pt

Trang 17

Thus, there are some essential reasons that make Lidar overcomes successfully

those limitation compared to other satellites: it provides its 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 Lidar is able to

penetrate high optically thin cloud and profile a large fraction of the atmosphere as

well as retrievals AOD from near surface to total column which is the main key point

to solve the uncertainty of AOD - PM2.5 relationship problem Furthermore, the

aerosol compositions also provides in occurrence of multiple aerosol layers which is

are contributed into the strength of such AOD - PM2.5 correlations (Schaap et al.,

2008)

Indeed, the statistical regressions between AOD and in situ PM concentration

measurements can be strongly improved by both retrieved AOD near surface and

aerosol compositions from Lidar It is a vital role to conduct research “Investigation

on the spatial distribution of PM 2.5 by integrating Satellite image from 2013-2015”

1.2 Research’s questions

1 Based on CALIPSO data, how AOD from full column and near surface

layers affected AOD-PM2.5 correlation and aerosol particles?

2 Can near surface observations from CALIPSO be used as a better proxy

for PM2.5 concentration?

3 Can aerosol classification from CALIPSO improves AOD-PM2.5

relationship in the near surface? What areis the main sources of regional air pollution

sources?

1.3 The requirement

1 Acquiring AOD AERONET from ground measurement, AERONET

Formatted: Indent: First line: 1.27 cm, Line spacing: Multiple 1.9 li

Formatted: Indent: First line: 1.27 cm, Space After: 0 pt, Line spacing: Multiple 1.9 li

Formatted: Font: Italic, Font color: Auto

Formatted: Indent: First line: 1.27 cm, Space After: 0 pt, Line spacing: Multiple 1.9 li, Tab stops: 2.25 cm, Left

Trang 18

2 Retrieving AOD CALIPSO from near surface to the total column with

CALIPSO products

3 CollecGatheringCollecting PM2.5 data from five differentground based

stations: Guting, Tucheng, Zhonghe, Xindian and Banqiao

2.1 Ground based Measurements

2.1.1 Ground based Measurement - PM2.5

Health effects of PM2.5 have been derived from epidemiological cohort studies in a

variety of geographical (principally urban) locations, mostly in the USA (Zheng, Pozzer,Cao,

& Lelieveldet al., 2015)

Exposure to fine particulate matter with aerodynamic diameters less than 2.5

µm (PM2.5) has negative effects on human health and may induce respiratory

problems, cardiovascular and lung diseases, and additional health problems (Pope et

al., 2002) They can be transported over longer distances than primary pollutants and

thus have a much broader spatial distribution than primary PM concentration (Ying &,

Q., Kleeman, 2009) Both short-term and long-term exposures to PM2.5 have been

linked to increased morbidity (Brunekreef,B., & Holgate, 2002) To get a better size

perspective and to understand how tiny those particles are a size comparison is given

in figure 1

Formatted: Line spacing: Double, No bullets

or numbering

Formatted: Font: 13 pt, Font color: Auto

Formatted: Font: 13 pt, Not Bold, Font color: Auto

Formatted: Indent: First line: 1.27 cm, Space After: 0 pt

Trang 19

Figure 1 1 : Size comparison between two aerosols with diameters 2.5 and 10 µm, a

human hair and a sand grain (credit: Environmental Protection Agency)

Although ground-based measurements are generally considered to be accurate,

they are representative for only relatively small areas around point stations Often, the

limited spatial coverage and irregular distribution of ground-based monitoring stations

largely restrict the study on space time dynamics of air pollution and its impacts on

human health and the environment Alternatively, complex process- based air pollution

models, which estimate pollutant concentrations by considering pollutant generation,

transportation, and removal, are hampered in a lot of cases by the incomplete

information of anthropogenic emission inventories and natural sources (Koelemeijer et

al., 2006) Thus, it is vital to access the source of air pollutants in various altitudes to

cast doubt on the local emission or long range transport cause

For the sake of monitoring and supervising the air quality in Taiwan, ground

PM2.5 measurement has established by Environmental Protection Administration

(EPA) Taiwan based on UK Daily Air Quality Index (DAQI) on 10/1/2014 with more

than 72 monitoring stations collects the measurements of PM10, PM2.5, SO2, NO2, O3,

Trang 20

and CO in real time In this study, hourly averaged PM2.5 concentrations when the

satellite passeds over the study region were were downloaded from the national air

quality publishing platform (http://taqm.epa.gov.tw/) PM2.5 value can be divided into

4 values with its corresponding effect to human health When PM2.5 index is 1 or 2,

at-risk individuals should reduce strenuous physical activity and particularly outdoors

When PM2.5 index is 3, general population should consider reducing activities and

outside When it reaches to level 4 that mean air quality is really bad and people

should reduce outdoor activities The PM2.5 concentration related to air pollution

banding L, M, H and VH represent low, moderate, high and very high, respectively

Reduce physical exertion, particularly outdoors, especially if you experience symptoms such as cough or sore throat

(Source: taqm.epa.gov.tw)

2.1.2 Ground based measurement - AERONET

The ground-based AERONET ( http://aeronet.gsfc.nasa.gov/) network of

Sun-sky radiometers (Holben , B N., et al., 2001) produces measurements of solar

direct-beam transmission and sky radiance that are inverted to yield aerosol column size

distributions and complex refractive indices at four wavelengths: 440, 670, 870, and

Formatted: No underline, Font color: Auto

Formatted: Font: Times New Roman, 13 pt

Formatted: No underline, Font color: Auto

Trang 21

1020 nm (Dubovik & King, 2000) have demonstrated the use of AERONET-retrieved

complex refractive index to derive information about both aerosol black carbon

content and water uptake AERONET provides columnar aerosol optical depths over

both land and ocean but is restricted to point observations (Alam et al., 2010) The

AOD observations are obtained from the AERONET program, which is a federation of

ground-based remote sensing aerosol networks to measure aerosol optical properties

(Holben et al., 1998) AERONET measurements can enable more accurate retrieval of

aerosol properties without surface reflectance information as it’s observe from direct

solar measurements AOD is a unit less optical representation of the column loading of

atmospheric aerosols The AERONET data provides AOD in the form of all points,

daily averages, and monthly averages The usefulness of AERONET retrieved the

abundant of aerosol optical depth (AOD), as indicator of aerosol composition,

including black carbon, organic matter, and mineral dust (Schuster G.,et al., 2005) to

validate with aerosol classification retrieved from AOD CALIPSO

2.2 Brief Description of Remote Sensing – CALIPSO

The Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations

(CALIPSO) mission was developed as part of the National Aeronautics and Space

Administration (NASA) Earth System Science Pathfinder (ESSP) program in

collaboration with Centre National d’E´tudes Spatiales (CNES), the French space

agency, with the goal of filling existing gaps in our ability to observe the global

distribution and properties of aerosols and clouds (Winker, 2003)

Lidar is the only technique giving high resolution profiles of aerosols, generates

vertically resolved distributions of aerosol types and their respective optical

Trang 22

characteristics which have significant contributions to the top-of-atmosphere radiation

(Omar et al., 2009) and it is able to observe aerosol above bright surfaces, such as

deserts and snow, and above bright clouds Because Lidar provides its 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

Furthermore, Lidar is able to penetrate high optically thin cloud and profile a large

fraction of the atmosphere There are also limitations in current cloud ice–water phase

retrievals from passive satellite sensors CALIPSO provides a vertically resolved

measurement of ice–water phase through measurements of the depolarization of the

Lidar backscatter signal So the side-scattering Lidar is very suitable to detect aerosol

spatial distribution in the boundary layer from the surface Three types of profiles are

provided in the level 2 products: total backscatter (parallel plus perpendicular) at

532nm and 1064nm and the 532nm perpendicular backscatter Vertical structure of the

atmosphere most aerosols are present in the lowest 1 to 2 km of the atmosphere,

particularly in the mixing layer However, it is not uncommon that substantial

aerosol-loaded air masses are present above the mixing layer These aerosols are decoupled

from the ground and usually originate from sources far away They are, therefore,

likely to have different (optical) properties than aerosols at ground level Thus, in-situ

measurements of aerosol properties at ground level are not always representative of

aerosol particles aloft and the total aerosol column above the measurement site To be

able to recognize such cases, the vertical structure of the atmosphere needs to be

known, and particularly about the presence of aerosol layers and clouds Lidar (Light

Detection and Ranging) instruments are well suited to detect aerosol layers, even

Comment [A1]: Xem lại Formatted: Font color: Auto

Trang 23

The orbit track locations of CALIPSO passes passed to Taiwan one day at

particular time (Figure 2)

Figure 2 2 : The orbit track of CALIPSO passe ds to Taiwan at 17:40pm on August 10,

2014 (Source: www-calipso.larc.nasa.gov)

As shown in figure 3, the signal strength has been color coded such that blues

correspond to molecular scattering and weak aerosol scattering; aerosols generally

show up as yellow/red/orange Stronger cloud signals are plotted in gray scales, while

weaker cloud returns are similar in strength to strong aerosol returns and coded in

yellows and reds At higher altitudes, the horizontal atmospheric structure is more

homogeneous, at lower altitudes, observation of profile is more likely heterogeneous

due to local effects (Held et al., 2012) Engel-Cox , et al., (2004) and He, et al., (2006)

pointed out that aerosol vertical profiles derived from Lidar observations could

Formatted: Font color: Auto, Expanded by 0.4 pt

Trang 24

improve the correlation between columnar AOD and surface measurements of PM or

extinction (Ffigure 3)

Figure 3 3 : Total Attenuated Backscattering signal measured by the CALIOP passed

Taiwan (red box) level 2 at 532 nm during the period 17:40- 17:54 UTC (Source:

www-calipso.larc.nasa.gov)

Lidar (Light Detection and Ranging) instruments are well suited to detect

aerosol layers, even above the mixing layer which is important parameter for

understanding the transport process in the troposphere, air pollution, weather and

climate change (Wang & Wang, 2014) In this study, Lidar provides information

on the vertical structure of the aerosol profile, atmospheric layering and the

presence of clouds up to an altitude of 15 km The backscatter Lidar operates at a

single wavelength (532 nm) has its ability to estimate aerosol optical properties,

in additionally to the qualitative vertical aerosol, layer location (both vertically

and horizontally) and cloud profile (Schaap et al., 2008) (Figure 4)

Aerosols Clouds

Formatted: Font color: Auto, Expanded by 0.3

Trang 25

Figure 4 4 : Vertical Feature Mask measured by the CALIOP passed Taiwan (red box)

level 2 during the period 17:40- 17:54 UTC (Source: www-calipso.larc.nasa.gov)

2.3 The Aerosol particles and The Normalized Gradient Aerosol Index (NGAI)

2.3.1 The role of aerosol types

Aerosol types have complex properties such as tiny size particle (micro and

submicron) and different in phases (liquid or solid) that are suspended in the

atmosphere (Balarabe et al., Abdullah, & Nawawi, 2016) (Balarabe, et al., 2016)

Depending upon their shapes, sizes and composition they can reflect sunlight back to

space and cool the atmosphere, they can also absorb sunlight and warm the

atmosphere Aerosols can even change the lifetimes of clouds, how much rainfall can

occur, and how they reflect sunlight They further can enable chemical reactions to

occur on Because cloud droplets form on aerosol particles, changes in aerosol

concentration and properties can alter cloud properties and precipitation There are

many sources of aerosols both natural and resulting from human activities with widely

varying distribution and properties Mineral dust consists of large, non-spherical

particles that absorb UV radiation due mainly to their iron oxide content Fresh smoke

Trang 26

from forest, agricultural, or grassland fires mainly consists of small particles that

absorb light in the UV and visible range (Dubovik, et al., 2002)

Moreover, most studies have concluded that the PM2.5-AOD relationship alone

cannot be used to estimate surface level PM2.5 since the vertical distribution of

aerosols and other meteorological parameters such as humidity and aerosol

compositions could also be important (Liu, et al., 2005; Paciorek, et al., 2008) Since

the Lidar measurements provide columnar retrievals, relating to this surface values

requires information about the vertical distribution of aerosols and aerosol

compositions

2.3.2 The Normalized Gradient Aerosol Index (NGAI)

Some of the aforementioned studies classify aerosol types by aerosol optical

depth (AOD) and Angstrom exponent (AE) These methods sort aerosol types into

dust (high AOD, low AE), marine (low AOD), and anthropogenic aerosols (high

AOD, high AE), but cannot subcategorize anthropogenic aerosols into absorbing and

non-absorbing without referring to geo-location information (Lee et al., 2010) The

category of atmospheric aerosols can be efficiently identified with the characteristics

of optical properties, such as Ångström exponent (AE) and single scattering albedo

(SSA) However, SSA can be not always retrieved from AERONET as well as

MODIS product so it caused the limitation in order to identify aerosol types (Lin et al.,

2016Wei, 2016)

In this study, the characteristics of aerosol types from AERONET is analyzed

on a global scale to determine the dominant aerosol type at each location using NGAI

Formatted: Font color: Auto, Not Highlight

Trang 27

burning and Anthropogenic pollutant, especially it has been successfully classified mix

- aerosols (dust and biomass burning, biomass burning and anthropogenic pollutant) as

shown in table 2

The extinction of sunlight by aerosols when it passes vertically through the

atmosphere from the top of the atmosphere to the surface is called the aerosol optical

depth (τa) In general, the optical properties of aerosols are the functions of

wavelength, and the spectral variance in radiometry may be inconsistent between types

of aerosols The general formula of NGAI is (Lin et al., 2016),

ref

NGAI τ(λ,λ ) / τλ

2 1

∇

=

2 1 ) , (

2 1 2

τ τ

When: τλref is the spectral AOD for aerosol loading normalization which can

be either , or another one are two wavelengths and are the

spectral AOD from two wavelengths

In this study, aerosols types are classified using blue band

(0.47µm), red band (0.66µm) and at 1.02 µm based on the slopes

of linear regression between multi-wavelengths and AOD (Lin et al., 2016Wei, 2016)

The AOD is only retrieved for cloud-free pixels and over surfaces that are not too

reflective The surface reflectivity at visible wavelengths is then obtained by assuming

a constant ratio between surface reflectivity at 1.01 µm and that at 0.47 and 0.66 µm

Levy et al (2007) also found indicated that by using higher surface reflectivity at 0.47

Trang 28

and 0.66 µm, the lowered AOD over land significantly reduced the discontinuity

between land and sea over the coastline in the north-east of the USA

Table 2: Aerosol classification using NGAI algorithm

2.3.3 AOD fraction of mixed type aerosols

NGAI value outside the intervals is initially determined to the situation caused

from mixed or other aerosol type When aerosol types are classified in wide area, the

higher mixture types may occur

Thus, the NGAI value from mixed type pixel would locate around the boundary

regions of intervals between aerosol types, known as the overlapping regions (Lin et

al., 2016Wei, 2016) AOD fraction of mixed type aerosols is detected by using the

concept of linear composite For the case of dual-type aerosols (type A and B), the

AOD fraction of type A and B can be derived from NGAI as below,

B mean A

mean

mixed A

mean A

AOD

NGAI NGAI

NGAI are the mean values of NGAI of aerosol type A and B, whileNGAI mixed is

the NGAI value of mixed type aerosols As shown in Figure 5 depicts the

approach of AOD fraction determination for dual-type aerosols (type A and B)

Formatted: Font: 1 pt, Font color: Auto

Formatted: Font: 13 pt, Expanded by 0.3 pt

Formatted: Font: 13 pt, Font color: Auto, Expanded by 0.3 pt

Trang 29

Figure 5 SEQ Figure \* ARABIC Error! No sequence specified. : The scheme of

AOD fraction determination for dual-type aerosols (type A and B) based on NGAI

values ( Lin et al.Wei , 2016)

Three kinds of dual-type aerosols are classified including AP mixed with

BB aerosols (APBB), Dust mixed with BB aerosols (DustBB) and Dust mixed with

AP aerosols (DustAP)

Formatted: Font: 13 pt, Italic, Font color: Auto

Trang 30

Description of Study aA rea

2.3.4 Study area

The locations of PM2.5 have been selected carefully to match with AERONET

site and CALIPSO track, located in Taipei city and new Taipei which is the capital city

of Taiwan Situated at the northern tip of Taiwan, Taipei City is an enclave of the

municipality of New Taipei However, with the rapidly development of the city, a

number of some environmental issues has have begun to emerge With the

additionAdditionally, of the significant features of the climate (cloudy and mist, short

sunshine is short and , high air humidity) and air pollution situationllutants in the city

which cannot rapidly diffuse rapidly, but, can be easy to accumulate in the city and

sub-urban Nevertheless, the dust from China due to long range transport is transferred

and affected much more to Taipei’s air quality

2.3.5 Software

Matlab and excel software have beenwere used to handle and analyze all

satellite aerosols, classified aerosol types as well as visualize the linear relationship

between AOD-PM2.5

Formatted: Font color: Red

Formatted: Font color: Black

Formatted: Font color: Red

Định dạng
Số trang	61
Dung lượng	1,51 MB