Object orientate application development II (III)

sd.docx

TABLE OF CONTENTS

Executive Summary

Statement of the problem

Aerosols and Human Health

Monitoring Particulate Matter Pollution

Previous Work

Data and Methods

MOD04 and MYD04

PM2.5 Mass Concentration

RUC20 Meteorological Fields

Ancillary Data Sets

Proposed Methods

Relevance and Outcomes

References

Executive Summary

The major goal of this proposal is to build an integrated analysis system using satellite and

ground-based aerosol datasets and model derived meteorological fields to estimate and monitor

particulate matter air quality over several locations in the United States and mega cities of the

world. This study will enable air quality assessment in remote areas where ground based

pollution measurements are not available. We propose to develop air quality indices over large

spatial scales that are not available currently because ground based information cannot provide

adequate coverage. This proposal will utilize NASA’s new space based sensors such as MODIS,

MISR, and CALIPSO (see acronym descriptions in Appendix A) that have tremendous potential

to support federal and state agencies involved in monitoring and forecasting of air quality. This

study will also explore the use of GEOS-CHEM model derived vertical information to build the

air quality system. Satellite remote sensing of surface level pollution is an innovative and a

promising area of research, which has immediate practical applications and possible long-term

benefits.

1. Statement of the problem

Urban air quality has gained critical public health concern in many parts of the globe as

urbanization and industrialization have amplified many folds during the last few decades.

Almost, half of the world’s population now lives in the urban areas and their number will

increase to four billion by the end of this decade. Particulate matter (PM) (or aerosols) and ozone

are the two major pollutants affecting the air quality in urban areas of the United States (US) and

throughout the world. Particulate matter is a complex mixture of solid and liquid particles that

vary in size and composition and remain suspended in the air. Many chemical, physical, and

biological components of atmospheric aerosols are identified as being potentially harmful to

respiratory and cardiopulmonary human health effects. Aerosols have many sources from both

natural and anthropogenic activities, naturally occurring processes such wind blown dust and

episodic activities such as forest fires/agricultural burning (mostly anthropogenic), dust storms

and volcanic eruptions. Increasing human activities also contribute to combustion from

automobiles, industries and emission from power plants. Apart from direct emissions, PM is also

produced by other processes such as gas to particle conversion in the atmosphere.

1.1 Aerosols and Human Health

Atmospheric aerosols are one of the most important components of the earth-atmosphere

system and play important role in climate and weather related processes [ Kaufman et al., 2002

and Ramanathan et al., 2001]. Air pollution has both short-term and long-term effects. Short

term impacts include, respiratory infections, irritation to the eyes, nose and throat, headaches,

nausea, and allergic reactions. Short-term air pollution can intensify the medical conditions of

individuals with asthma and emphysema. In 1952 London experienced one of the worst smog

disasters, which killed more than four thousand people in few days due to very high

concentration of particulate matter in the air [ Scarrow, 1972]. Long-term effects include lung

cancer, heart disease, chronic respiratory disease, and even damage to the brain, nerves, liver, or

kidneys. Continual contact to air pollution affects the lungs of growing children and may worsen

or complicate medical conditions in the elderly.

Particulate matter with aerodynamic diameters less than 2.5 µm (PM2.5) can cause respiratory

and lung disease and even premature death [ Krewski et al., 2000]. The World Health

Organization (WHO) estimates that 4.6 million people die each year from causes directly

attributable to air pollution. Worldwide more deaths per year are linked to air pollution than to

automobile accidents. Some examples from all around the world: approximately 310,000

Europeans die from air pollution annually [ van Leeuwen van, 2002], The Tata Energy Research

Institute (TERI) in India estimated 18,600 premature deaths per year associated with poor air

quality in the Delhi region [ TERI, 2001], increased PM was associated with 2400 deaths per year

in Australia with an associated health cost of $17.2 billion [ Morgan et al., 1998 , Simpson et al.,

2000] and Sydney experiences around 400 premature mortalities each year due to increased

levels of pollution, and asthma is also common in this area [ Barusch, 1997]. Similar mortality

deaths are associated with air pollution in other parts of the world. Direct causes of air pollution

related deaths include aggravated asthma, bronchitis, emphysema, lung and heart diseases, and

respiratory allergies. A medical study by [ Pope III et al., 2002] concludes that fine particles and

sulfur oxide related pollution are associated with all-cause, lung cancer and cardiopulmonary

mortality. The same study also states that an increase of 10 µgm-3 in fine particulate can cause

approximately a 4%, 6% and 8% increased risk of all cause, cardiopulmonary, and lung cancer

mortality, respectively. Using statistical data collected in twenty big cities, Samet et al. [2000]

showed that the daily mortality within a metropolitan area is associated with concurrent or

lagged daily fluctuations in ambient PM concentrations. Apart from impact on human health,

poor air quality also affects the health of animals and plants. Poor air quality conditions are also

associated with damaging buildings and monuments around the world. Indirectly air pollution

significantly affects the economy by increasing medical expenditures and expenditure for

preserving the surrounding environment.

1.2 Monitoring Particulate Matter Pollution

The US Environment Protection Agency (EPA) monitors air quality by measuring PM and

ozone concentration at thousands of ground based monitoring stations across the country. The

PM2.5 is measured using a Tapered-Element Oscillating Microbalance (TEOM) instrument with

an accuracy of ±1.5 µgm-3 for hourly averages. TEOM first collects the particles (<2.5 µm in

diameter) on Teflon coated glass fiber filter surface by passing them through a cyclone inlet,

which removes the bigger size particles from the sample of air. The inlet is heated to 50°C prior

to particles being deposited onto the filter in order to eliminate the effect of condensation or

evaporation of particle water. The filter is attached to a vibrating hollow tapered glass tube. In

the mass transducer unit, as the filter progressively load PM2.5 particles, oscillation frequency of

glass tube changes proportionally. The change in frequency of oscillation is directly related to

the mass of particles on element (filter), which can be measure using computer controlled unit

and hence the mass of PM2.5 is obtained in the unit of µgm-3 [ Charron et al., 2004].

The United States Environmental Agency (EPA) issues National Ambient Air Quality

Standards (NAAQS) for six criteria pollutants namely ozone, particulate matter, carbon

monoxide, sulfur dioxide, lead and nitrogen oxides. Standards for particulate matter were first

issued in 1971 then revised in 1987 and 1997 by EPA. Recently (September 2006), EPA revised

1997 standards to tighten the criteria. The 2006 standards reduced the 24-hour mean PM2.5 mass

concentration standard from 65 µgm-3 to 35 µgm-3, and retained the current annual PM2.5

standard at 15 µgm-3. The EPA reports an Air

Quality Index (AQI) based on the ratio between 24-hour averages of the measured dry particulate

mass and NAAQS, and it can range from nearly zero in a very clean atmosphere to 500 in very

hazy condition. Table 1 gives details on PM2.5 mass, air quality categories and possible health

effects. Currently USEPA provides particulate matter air quality forecast over more than 200

cities on daily basis. In recent years, other countries in Europe, Australia, Japan, and China have

also started monitoring PM2.5 mass as measure of air quality conditions. However these EPA and

other agencies monitoring stations are only point locations and do not have the spatial resolution

to map the regional to global distributions of aerosols.

Satellite data have tremendous potential for mapping the global distribution of aerosols and

their properties [ Chu et al., 2002]. However, several outstanding issues remain in using satellite

data because most satellite data provide column information whereas air pollution near the

ground is the most important parameter affecting human health. Several studies have

demonstrated the potential of monitoring air quality using high resolution data from space based

sensors over regional to global scales. The next section provides summary and key conclusions

from these studies.

2. Previous Work

Satellite remote sensing of particulate matter (PM) air quality is a relatively new area of

research in the field of atmospheric science. As presented in Table 2, many research studies have

shown the potential of using satellite derived aerosol optical thickness information as surrogate

for air quality conditions. Table 2 presents relevant published research in this field. The two

main conclusions from Table 2 are very clear; 1) Most of studies have used MODIS derived

AOT products except a few studies by Liu et al. [2004, 2005, 2007] , and van Donkelaar et al,

[2006] , which used AOTs from both MISR and MODIS. 2) Area of study for most of the studies

have been in some part of United States except studies by Gupta et al., 2006, Koelemeijer et al.

[2006] and van Donkelaar et al. [2006] . One of the reasons is that MODIS gives much better

spatial and temporal coverage as compared to MISR and measurements of PM2.5 mass

concentration in other parts of the world are limited. The first study by Wang and Christopher

[2003] used PM2.5 mass and MODIS AOT data over seven stations in Alabama and presented

very good correlation (>0.7) between these two parameters. This study also concluded that

although deriving exact PM2.5 mass from satellites could have larger error; satellites can provide

daily air quality condition as calculated by EPA with sufficient accuracies. It also shows the

potential of satellite monitoring of transport of air pollution from source to near and far urban

areas. Hutchison et al. [2004, 2005] , mainly focus on air quality over Texas and Eastern United

States and use of satellite imaginary in detecting and tracing the pollution. The first

comprehensive study by Engel-Cox, et al. [2004] presented a thorough correlation analysis

between MODIS AOT and PM2.5 mass over entire United States. The correlation pattern shows

high values in Eastern and Midwest portion of the United States whereas correlations are low in

Western United States. The authors also states that ‘ This variability is likely due to a

combination of the differences between ground-based and column average datasets, regression

artifacts, variability of terrain, and MODIS cloud mask and aerosol optical depth algorithms.’

This study also concludes that high space and time resolved observations from satellites can

provide synoptic information, visualization of the pollution, and validation of ground based air

quality data and estimations from models. Engel-Cox and other co-authors also published other

studies in 2004, 2005 and 2006 which further emphasizes the use of satellite derived aerosol

products in day to day air quality monitoring and even in policy related decision making. One of

these papers [ Engel-Cox, et al., 2006] also presented the application of LIDAR derived vertical

aerosol profiles to improve PM2.5 -AOT relationship. MODIS aerosols and clouds data are now

being used in the IDEA (Infusing satellite Data into Environmental Applications) program to

monitor air quality over United States. IDEA is a joint effort from NASA, NOAA and EPA to

improve air quality assessment, management, and prediction by infusing satellite measurements

into analysis for public benefit [ Al-Saadi et al., 2005]. MISR derived aerosol products were first

used by Liu et al. [ 2004] , which shows similar potential for air quality applications. This study

also used GOCART and GEOS-CHEM models derived meteorological fields to examine their

relative effects on PM2.5 -AOT relationships. Liu et al. [ 2007] also compared MISR and MODIS

over few sites around St. Louis area, which shows better performance of MISR than MODIS for

air quality application. Gupta et al. [ 2006] compared PM2.5 -AOT relationship in different parts

of the world such as Europe, Australia, USA, and Asia. This study shows applications of satellite

derived air quality products at global scales and in the regions where surface PM2.5

measurements are not available. Correlations analysis varies in different parts of the world

depending on accuracies of MODIS retrieval, cloud contamination in AOT and height of aerosol

layer in the atmosphere. Results from this study are presented in preliminary result section. van

Donkelaar et al, [2006] used GEOS-CHEM derived vertical extinction profiles and basic mass

formula to calculate mass of fine particles and compared the results over several locations in

USA and Canada. This study also presented global picture of MODIS and MISR derived PM2.5

mass concentration and results looks more promising.

All these studies mainly concluded that the MODIS and MISR AOTs are important to define

air quality over large spatial domains and to track and monitor aerosol sources and transport.

These studies are based on correlation and linear and multi-variant regression between MODIS

AOT, ground based PM2.5 mass and model derived meteorological parameters. The MODIS

derived AOT which is measure of column aerosol loading cannot be used alone to derive PM2.5

mass concentration, which is an indicator of the mass of the dry PM2.5 near the surface [ Wang

and Christopher, 2003] . Meteorological factors such as surface temperature (Ts), relative

humidity (rh), wind speed (WS) and direction (WD), variations in sunlight due to clouds and

seasons are important parameters which affect the relationship between the two parameters.

Changes in these processes, which affects the variability in pollution, is primarily governed by

the movement of large-scale high and low-pressure systems, the diurnal heating and cooling

cycle, and local and regional topography. The vertical profile of aerosol mass extinction ( ext),

which determines effective scale height (Heff) and hygroscopic growth factor (a function of rh)

are also very important parameters that must be accounted for while deriving relationships

between PM2.5 and AOT [ Wang & Christopher, 2003, Gupta et al., 2006). Strong winds of 6 ms-

1 or more can cause dust to become airborne and many factors influence the amount of PM2.5

produced by windblown dust including vegetation cover, soil moisture, soil particle size

distribution, surface roughness, and changes in wind direction ( Saxton et al., 2000). Easterly

trade winds can transport Saharan dust to the eastern and southeastern US [ Prospero, 1999] and

can increase PM2.5 concentrations at the surface and degrade visibility. Also the western US can

be affected by dust transported from Asia [ Falke et al., 2001]. Air quality modeling actually

requires a system of models and observations including satellite and ground-based data that work

together to simulate the emission, transport, diffusion, transformation, and removal of air

pollution and these models include meteorological models, emission models and air quality

models.

4. Data and Methods

This section will list and describe different data sets and the methods, which will be used for

research.

4.1 MOD04 and MYD04

The Moderate Resolution Imaging Spectroradiometer (MODIS) onboard NASA’s Terra

(morning satellite with equatorial over cross time is 10:30 AM) and Aqua (afternoon satellite

with equatorial over cross time is 1:30 PM) satellites give systematic retrieval of cloud and

aerosol properties over land [ King et al., 1999]. MODIS provides the spectral information on

aerosol optical properties in seven different wavelengths with good accuracy [ Kaufman et al.,

1997, Remer et al., 2002 and 2005]. Aerosol optical thickness (AOT) is an important aerosol

parameter retrieved from satellite observations representing columnar loading of aerosols in the

atmosphere along with the fraction of fine mode aerosol which is an indicator of anthropogenic

pollution which correlates well with PM2.5 mass. Several validation studies conducted over land

reveal that 57% MODIS AOT retrievals are within expected uncertainty levels of ± 0.05 ± 0.15

AOT [ Remer et al., 2005]. Preliminary results from validation exercise of MODIS collection 5

shows increase in the number from 57% to 67%. Depending on availability, MISR derived AOT

will also be used to inter-compare the results from two different sensors. As an example Figure 1

presents the aerosol optical thickness observed by MODIS (Terra) and estimated air quality

conditions during August 20-26, 2006 over EPA region 4.

4.2 PM2.5 Mass Concentration

The United Stated Environment Protection Agency and its state partners maintain several air

quality monitoring networks in the United States. The networks monitor the mass concentration

and speciation (some of the sites) of gaseous and particulate air pollutants at the ground level.

PM2.5 data from these networks include 24-h average (daily) concentration data, typically taken

every 3 days, and continuous (hourly) PM2.5 concentration measurements. Most relevant for

present study is PM2.5 mass concentration measure in µgm-3. This study will use both hourly and

daily mean PM2.5 mass data sets from several ground stations in different area of interest as

shown in figure 4 .

4.3 RUC20 Meteorological Fields

The Rapid Update Cycle (RUC) is an operational atmospheric prediction system comprised

primarily of a numerical forecast model and an analysis system to initialize that model. The RUC

has been developed to serve users needing short-range weather forecasts. RUC runs

operationally at the National Centers for Environmental Prediction (NCEP). A new version of

the RUC has been implemented at the NCEP with a improved horizontal resolution (20km),

increased number of vertical computational levels (40 to 50), and improvements in the analysis

and model physical parameterizations. A primary goal in development of the 20-km RUC (or

RUC20) has been improvement in warm-season and cold-season quantitative precipitation

forecasts. Improvements in near-surface forecasts and cloud forecasts have also been targeted.

The RUC20 provides improved forecasts for these variables, as well as for wind, temperature,

and moisture above the surface [ Benjamin et al., 2002]. Hourly data of relative humidity, wind

speed, wind direction, and height of planetary boundary layer (PBL) from RUC20 will be used in

the proposed study.

4.4 Ancillary Data Sets

In order to build air quality model several supporting data sets will be required These include

chemical speciation from IMPROVE network, GEOS-CHEM derived vertical aerosol layer

heights and CALIPSO (for case studies in 2006-2007 time periods) derived vertical aerosol

optical thickness profiles. To extend the study to other mega cities of the world, similar data sets

for all other cities will be collected. For this study, all the cities, where human population is more

than ten million are categorized as megacities. Air quality in such cities is mainly governed by

anthropogenic activities due to large human population. 20 such megacities are already identified

and relevant data sets for each city will be collected.

4.5 Proposed Methods

Initially, the proposed study will use data sets available during year 2005 over different

months and seasons. This study will primarily focus on EPA region 4 but other areas as shown in

figure 4(a) will also be explored depending on availability of different data sets and time line of

the current project. These study areas are selected due to complexity of aerosols (e.g. dust from

Saharan, smoke from Central America, anthropogenic urban pollution, emissions from power

plants and the ease of data access. Also, it will be important to monitor the accuracies of satellite

derived products in different part of the study area. Mega cities (population greater than 4

million) analysis will be done for three year time period starting January 2003 to December

2005.

First we will obtain all data sets and apply quality control processes to all the data sets. For

example, MODIS AOT in valid ranges with ‘Good’ quality flag will be used for only clear sky

conditions using reported cloud cover percentages. Different statistical parameters such as mean,

median, and standard deviation will also be calculated and by applying different conditions on

these values for each data set, qualified data sets will be selected. Using appropriate weighting

functions, the MODIS AOT will be collocated in space and time with ground based PM2.5

measurements [ Gupta et al., 2006]. All the PM2.5 mass measurements and other meteorological

parameters will be averaged within one hour of MODIS observation time. Since RUC derived

meteorological fields are available in 20 km horizontal resolution hence a box of 20X20 km

centered at air quality stations will be used to average AOT, and meteorological fields. To test

the sensitivity of vertical profiles of aerosols, GEOS-CHEM derived coarse resolution monthly

means data will be used. Depending on availability, LIDAR profiles as well as observations from

CALIPSO will be used. Once quality controlled data sets of hourly averaged PM2.5 mass (µgm-

3), MODIS AOT (0.55 µm), meteorological parameters such as wind speed, wind direction,

relative humidity, and height of planetary boundary layer are ready, the training of ANN system

will start.

Several case studies with extreme pollution events such as dust storm and forest fires will be

first identified (e.g. the transport of aerosols from biomass burning fires in Central America to

South Eastern United States) for analysis. Variability and accuracy of the output PM2.5 mass will

be tested for different seasons at different geographical locations. The sensitivity studies will be

performed to study the role of each input parameter in the model. The model will also be tested

against extreme values of input parameters such as very high wind speed, rainy conditions, and

extreme low temperatures. Errors will be reported as root mean square values during validation

exercise between measured and estimated PM2.5 mass concentration. In order to minimize these

errors, different exercises will be carried out on the input data sets. This may includes putting

constraints on AOT, PM2.5 and meteorological fields.

8. Relevance and Outcomes

The major goal of this proposal is to build an integrated analysis system using satellite and

ground-based datasets and meteorological data to monitor and predict particulate matter air

quality over the several locations in the United States and mega cities of the world. Satellite

remote sensing of surface level pollution is an innovative and promising area of research, which

has immediate practical applications and long-term benefits. The expected results from this

research will enhance our scientific understanding on satellite remote sensing of particulate

matter and can be utilized toward operational monitoring of particulate matter air quality over

remote areas where point observations are not available.

References

· Al-Saadi, J., et al. (2005), Improving National Air Quality Forecasts with Satellite Aerosol

· Observations, Bulletin of the American Meteorological Society, 86(9), 1249-1261.

· Barusch, R. (1997), Air pollution and health, In AMA Forum on Air Pollution and Health.

· Benjamin, S., G., et al. (2002), RUC20: The 20-km version of the Rapit Update Cycle, NWS

· Technical Procedures Bulletin, vol 490.

· Charron, A., R. M. Harrison, S. Moorcroft, and J. Booker (2004), Quantitative interpretation of

· divergence between PM10 and PM2.5 mass measurement by TEOM and gravimetric (Partisol)

· instruments, Atmospheric Environment, 38(3), 415-423.

· Chu, D. A., Y. J. Kaufman, C. Ichoku, L. A. Remer, D. Tanré, and B. N. Holben (2002),

· Validation of MODIS aerosol optical depth retrieval over land, Geophysical Research Letters,

· 29(12), 8007.

· Engel-Cox, J. A., T. Hoffmann, and A. D. J. Haymet (2004), Recommendations on the use of

· satellite remote-sensing data for urban air quality, Journal of the Air Waste Management

· Association 54(11), 1360-1371.

· Engel-Cox, J. A., G. S. Young, and R. M. Hoff (2005), Application of satellite remote-sensing

· data for source analysis of fine particulate matter transport events, J Air Waste Manag Assoc,

· 55(9), 1389-1397.

· Engel-Cox, J. A., C. H. Holloman, B. W. Coutant, and R. M. Hoff (2004), Qualitative and

· quantitative evaluation of MODIS satellite sensor data for regional and urban scale air quality,

· Atmospheric Environment, 38(16), 2495-2509.

· Engel-Cox, J. A., R. M. Hoff, R. Rogers, F. Dimmick, A. C. Rush, J. J. Szykman, J. Al-Saadi, D.

· Chu, and E. R. Zell (2006), Integrating lidar and satellite optical depth with ambient

· monitoring for 3-dimensional particulate characterization, Atmospheric Environment, 40(40),

· 8056-8067.

10