8/10/2019 Air Quality Assessment over Gulf Coast
1/39
Venkata B. Dodla, and colleagues
Trent Lott Geospatial Visualization Research Centre
College of Science, Engineering &Technology
Jackson State University, Jackson MS 39217, USA
________________________________________________________
8thInternational Symposium on Recent Advances in Environmental Health Research
Jackson, MS, USA
September 20, 2011
Air Quality Assessment over Gulf Coast by Establishing
Relationship Between MODIS-Satellite Derived Aerosol
Optical Thickness (AOT) and Surface PM2.5 Observations
8/10/2019 Air Quality Assessment over Gulf Coast
2/39
AQI Forecasting /Ozone/PM 2.5
This presentation is on our studies AQI Forecasting for PM2.5using SatelliteRemote sensing Data (MODIS Data)
o To Provide year-round coverage
o EPA pilot effort focusing on 36 major cities across the U.S
o Currently developing PM2.5 forecast tools for State and Local Agencies MODIS/AIRNow data fusion would be one of many tools
1. Simulation of surface Ozone pollution in the Central Gulf Coastal region during
summer synoptic condition using WRF/Chem air quality model. 2011,
Atmospheric Pollution Research, DOI:10.5094/APR.2012.005
2. Air Quality Modeling for Urban Jackson, MS Region using High Resolution
WRF/Chem Model. 2011, Int. J. Environ. Res. Public Health, 8, 2470-2490;doi:10.3390/ijerph8062470.
3. Sensitivity Evaluation of the WRF/Chem to Different PBL and Land Surface Physics
for Air Quality Simulations in the Mississippi Gulf Coastal Region. 2010, Advances
in Meteorology, Volume 2010, Article ID 319138, 24 pages. doi:10.1155/2010/
319138, Hindawi Publishing Corporation, New York, USA.
Ozone and PM2.5 are high significant pollutants for Air Quality Index(AQI)
Our Group earlier published a number of papers on simulation of Ozone
levels in Gulf Coast
8/10/2019 Air Quality Assessment over Gulf Coast
3/39
1:18
Coordinated Observations with the A-Train
TES
T, P, H2O, O3, CH4, COMLSO3, H2O, CO
HIRDLST, O3, H2O, CO2, CH4
OMIO3, aerosol climatology
aerosols,
polarization
CloudSat3-D cloud
climatologyCALIPSO3-D aerosol
climatology
AIRST, P,H2O, CO2, CH4
MODIScloud,
aerosols, albedo
OCO- CO2O2A-band
ps, clouds,
aerosols
8/10/2019 Air Quality Assessment over Gulf Coast
4/39
8/10/2019 Air Quality Assessment over Gulf Coast
5/39
Air Pollution
Airborne particles and gases occurring in
concentrations that endanger the health andwell-being of organisms or disrupt the orderly
functioning of the environment. Lutgens and
Tarbuck
Air pollutants are categorized into two
categories:
Primary PollutantsSecondary Pollutants
8/10/2019 Air Quality Assessment over Gulf Coast
6/39
Primary Pollutants
Anthropogenic: Combustion Processes
Chemical Processes
Nuclear or Atomic Processes
Roasting, Heating and RefiningProcesses
Mining, Quarrying and Farming
Processes
Natural:
Volcanoes Breaking Seas
Pollens and Terpenes
Fire
Blowing Dust
Bacteria and Viruses
Primary Pollutants are airborne particles that are emitted directly fromidentifiable sources. These tiny structures are known collectively as Particulatematter (PM). Once suspended in either air or water, the mixture of the two
becomes known as an Aerosol.
8/10/2019 Air Quality Assessment over Gulf Coast
7/39
Secondary Pollutants
Secondary Pollutants are not emitted directly into the air, but form in theatmosphere from reactions taking place between primary pollutants.
SMOG (Smoke + Fog) VOG (Volcanic + Smog)
Ground Level Ozone
8/10/2019 Air Quality Assessment over Gulf Coast
8/39
OzoneStratospheric Ozone: GOOD Ozone
Contains 90% of atmospheric ozone
Primary Shield for UV Radiation from Sun
Produced by UV rays interacting with Oxygen
Tropospheric Ozone: Bad Ozone
Other 10% of Atmospheric Ozone
Forms close to the ground when
Hydrocarbons and Nitrogen Oxides react
with sunlight.
Detrimental to human respiratory health
Slows and alters growth of many species
of plants
8/10/2019 Air Quality Assessment over Gulf Coast
9/39
How are these pollutants spread?
Wind
Dilution is the solution topollution
Wind causes bodies of polluted air
to spread out across the Earth.
When winds are high, these
bodies are spread out over larger
areas, leading to a lower
concentration of pollutants.
When winds are low, these bodies
sit and stagnate over a smaller
area leading to localized pocketsof more concentrated pollution.
8/10/2019 Air Quality Assessment over Gulf Coast
10/39
Health Impacts of Air Pollution
8/10/2019 Air Quality Assessment over Gulf Coast
11/39
OBJECTIVES Establishing the relationship between the ground-level concentration of
fine particulate matter (PM2.5) using space-based measurements from the
Moderate Resolution Imaging Spectroradiometer (MODIS)
Improve accuracy of PM2.5AQI forecast for Regional/Seasonal
Characteristics and concentrations of air pollution is required at all
locations for assessment and implementation of mitigation measures.
prevent health and environmental impacts. Since monitoring at all
locations is not possible, improving observation network is essential,
utilization of satellite data improves spatial observations.
PM2.5 Observations MODIS-AOT Satellite
Establishing relationship between
PM2.5
and AOT
Estimation of PM2.5from AOT
8/10/2019 Air Quality Assessment over Gulf Coast
12/39
Percent increase in monthly mortality per increase in 1 g/m3of PM2.5
concentrations (June, 2000)
8/10/2019 Air Quality Assessment over Gulf Coast
13/39
Source: US EPA, 2003
PM2.5Continuous Monitoring Sites Reporting to AQS & AIRNow7/7/03
8/10/2019 Air Quality Assessment over Gulf Coast
14/39
Average quarterly PM2.5concentrations based on AIRNow data from 2005-2007
PM2.5concentrations
vary by
season and
region
Seasonal Patterns-National/Regional
8/10/2019 Air Quality Assessment over Gulf Coast
15/39
Derivation of Aerosol Optical Thickness (AOT)
from MODIS satellite data
Satellite imagery,
Large spatial coverage and reliable repeated measurements,
Advantageous tool to monitor aerosols and their transport patterns compared to
ground measurements
Aerosol Optical Thickness (AOT) - parameter - air quality.
AOT is the degree to which aerosols prevent the
transmission of light.
.
8/10/2019 Air Quality Assessment over Gulf Coast
16/39
Launched on Terra 12/18/1999, on Aqua 3/24/2002
Terra crosses Equator ~10:30 am, Aqua ~1:30 pm (orbit time is ~98
min)
36 spectral bands with resolutions spatial resolutions
250 m (bands 1 - 2)
500 m (bands 3 - 7) 1000 m (bands 836)
Global coverage 1-2 days
Swath Width ~ 2330 km
Altitude - 705 km
Orbit - Near-polar, Sun Synchronous
MODIS instrument on Terra satellite collecting
swaths of data over the Earth
MODIS (Moderate Resolution Imaging Spectroradiometer)
8/10/2019 Air Quality Assessment over Gulf Coast
17/39
Terra collects data on
descending node
Aqua collects data on
ascending node
TERRA
AQUA
8/10/2019 Air Quality Assessment over Gulf Coast
18/39
o The MODIS Aerosol Product (MOD04 (Terra), MYD04 (Aqua)) monitors the
ambient aerosol optical thickness over the oceans globally and over the
continents.o The aerosol size distribution is derived over the oceans, and the aerosol type
is derived over the continents. Fineaerosols (anthropogenic/pollution) and
courseaerosols (natural particles; e.g., dust) are also derived.
o The aerosol product includes the deep-bluealgorithm recently developed to
get aerosol optical thickness over bright land areas.
o Aerosol Total Optical Thickness is available through Giovanni at 550nm fromMODIS.
o Daily Level 2 AOT data available
o Time duration : Five minute
o Coverage: Once in a day
o Resolution: 10km X 10km - 1km pixels at nadir
o 3 spectral bands used in retrieving AOT - 0.47, 0.66 and 2.1 m
o GSFC urban/industrial aerosol model over Eastern United States used for
Retrieval of MODISAOT
Description of MODIS - Aerosol Product
8/10/2019 Air Quality Assessment over Gulf Coast
19/39
PM2.5 Surface Observations
o US EPA Monitors all over the USA.
Study region covers MS, parts of LA.
o ~32 monitoring sites in study region.
o PM2.5 data has been collected for
the entire year 2007 data with an
hour interval.
http://www.epa.gov/ttn/airs/airsaqs/
detaildata/downloadaqsdata.htm
Data used in the present study
MODIS AOT Datasets
AOT Data has been collected from both TERRA and AQUA for the year 2007 once in
a day, and the product is
Optical_Depth_Land_And_Ocean Description: Aerosol Optical Thickness at 0.55
m for both Ocean (best) and Land (corrected) with best quality data (QA
Confidence Flag = 3)
Dimensions: (Cell_Along_Swath, Cell_Across_Swath); (203,135)
Valid Range: -0.05 to 3.0 (Collection 050)
http://ladsftp.nascom.nasa.gov/allData/5/MOD04_L2/---- terra
http://ladsftp.nascom.nasa.gov/allData/5/MOD04_L2/ -- aqua
8/10/2019 Air Quality Assessment over Gulf Coast
20/39
As MODIS AOT data is available at 10kmx10km horizontal resolution; and surface
PM2.5 observations are available at point locations, we adopted two
methodologies to estimate the AOT value corresponding to the PM2.5 observationlocation.
Two kinds of PM2.5 datasets were prepared from hourly observations as (1) Daily-
Averaged value (2) MODIS Swath hour observation.
(1) MODIS-AOT at the four grid points (Grid Box method) around PM2.5
observation location are used to compute a simple and a distance weightedaverage
(2) Collect MODIS-AOT data falling within a specified radius and using them to
compute a simple average and a distance-weighted average. In this study we
used three values for radius as (i) 20 km; (ii) 50 km and (iii) 100 km
Both these methodologies are constrained to have at lease 4 data points.
TLGVRC developed the required FORTRAN software to retrieve the AOT data as an
in-house product.
Methodology
8/10/2019 Air Quality Assessment over Gulf Coast
21/39
Relationship between MODIS AOT and surface PM2.5
AOT (aerosol optical thickness)
MODIS (AQUA/TERRA) Data product
AOT vs PM2.5 relationship
Regression analysis
Software development
(TLGVRC)PM 2.5
Surface observations
Retrieval of AOT
to station location
Satellite Data has following Limitations
(i) AOT can be derived only during day time hence only one observation available per day.
(ii) AOT only available under less cloud cover or no-cloud.
(iii) Both in Grid Box method and Radial Distance Method, must have the AOT data at least
4 points.
8/10/2019 Air Quality Assessment over Gulf Coast
22/39
Averaged AOT from 4 neighboring points around PM2.5 observation site location given as
Distance Weighted AOT from 4 neighboring data points around PM2.5 Station location given as
4
y4)(x4,MODIS
y3)(x3,MODIS
y2)(x2,MODIS
y1)(x1,MODIS
),(
yxAOT
W4W3W2W1
W4*y4)(x4,
MODISW3*y3)(x3,
MODISW2*y2)(x2,
MODISW1*y1)(x1,
MODIS
),(.
yxAOTW
8/10/2019 Air Quality Assessment over Gulf Coast
23/39
8/10/2019 Air Quality Assessment over Gulf Coast
24/39
Averaged AOT computed within the defined radius of around the PM2.5 observation site
location given as
n
iyx
AOT1
ii
n
)y,(xMODIS),(
Distance Weighted Average AOT computed within the defined radius of around the PM2.5
observation site location given as
n
i
i
i
i
yx
W
W
AOTWeighted
1
n
1 ii
),(
*)y,(x
MODIS
_
n-total number of observationsavailable within the required radius
n-total number of observations available within the required radius
W-weight (here distance between observation site location and we consider as weight)
8/10/2019 Air Quality Assessment over Gulf Coast
25/39
PM2.5 VS AOT, obtained from MODIS-TERRA swath hour considering all
points within 100 km radial distance and for the year 2007 data.
No of observations
8/10/2019 Air Quality Assessment over Gulf Coast
26/39
Correlation is better with hourly values as compared to daily averages. There is
no significant difference between simple and weighted averages.
Time MethodTotal No. of
Data PointsCC MAE RMSE
Daily
Average
SA 149 0.13 5.48 9.42
WA 149 0.13 5.48 9.42
Swath
Hour
SA 161 0.31 6.15 7.45
WA 161 0.33 6.15 7.45
99.9% Statistical significance
SASimple Average; WA-Distance Weighted Average
CC-Correlation Coefficient; MAE-Mean Absolute Error
RMSE-Root Mean Square Error
Relationship between PM2.5 and AOT from TERRA SatelliteGrid Box Method
8/10/2019 Air Quality Assessment over Gulf Coast
27/39
Radius
(km)Time Method
Total No. of
Data PointsCC MAE RMSE
100
DailyAverage
SA 969 0.29 4.78 6.64
WA 969 0.29 4.79 6.65
Swath
Hour
SA 2041 0.31 5.91 8.59
WA 2041 0.30 5.92 8.62
50
Daily
Average
SA 864 0.28 4.83 6.76
WA 864 0.27 4.83 6.77
Swath
Hour
SA 1660 0.32 5.92 8.79
WA 1660 0.30 5.94 8.84
20
Daily
Average
SA 173 0.20 5.05 6.55
WA 173 0.19 5.06 6.57
Swath
Hour
SA 251 0.46 6.10 8.72
WA 251 0.43 6.13 8.88
99.9% Statistical significance
99% Statistical significance
SASimple Average; WA-Distance Weighted Average
CC-Correlation Coefficient; MAE-Mean Absolute Error
RMSE-Root Mean Square Error
Relationship between PM2.5 and AOT from TERRA SatelliteRadius Method
8/10/2019 Air Quality Assessment over Gulf Coast
28/39
Correlation is better with hourly values as compared to daily averages. There is
no significant difference between simple and weighted averages.
Time MethodTotal No. of
Data PointsCC MAE RMSE
Daily
Average
SA 7 -0.13 3.09 3.82
WA 7 -0.13 3.09 3.82
Swath
Hour
SA 24 0.60 3.60 4.54
WA 24 0.60 3.60 4.54
99.9% Statistical significance
SASimple Average; WA-Distance Weighted Average
CC-Correlation Coefficient; MAE-Mean Absolute Error
RMSE-Root Mean Square Error
Relationship between PM2.5 and AOT from AQUA SatelliteGrid Box Method
8/10/2019 Air Quality Assessment over Gulf Coast
29/39
Radius
(km)Time Method
Total No. of
Data PointsCC MAE RMSE
100
DailyAverage
SA 743 0.23 5.28 7.34
WA 743 0.23 5.28 7.34
Swath
Hour
SA 2037 0.30 5.64 7.88
WA 2037 0.30 5.64 7.89
50
Daily
Average
SA 645 0.22 5.24 7.43
WA 645 0.20 5.26 7.45
Swath
Hour
SA 1677 0.29 5.56 7.78
WA 1677 0.28 5.58 7.81
20
Daily
Average
SA 95 0.31 4.50 5.55
WA 95 0.33 4.51 5.50
Swath
Hour
SA 253 0.37 5.45 6.84
WA 253 0.38 5.45 6.83
99.9% Statistical significance
99% Statistical significance
SASimple Average; WA-Distance Weighted Average
CC-Correlation Coefficient; MAE-Mean Absolute Error
RMSE-Root Mean Square Error
Relationship between PM2.5 and AOT from AQUA Satellite
8/10/2019 Air Quality Assessment over Gulf Coast
30/39
Conclusions
MODIS AOT data has been collected over MS region and PM2.5 surface
observations from US EPA to establish relationship between them.
(i) A positive relation exists between PM2.5 and Aerosol Optical Thickness.
(ii) Correlations are high with the MODIS Pass time observations hour as
compared to the Daily averaged PM2.5 observations.
(iii) No significant differences found with Simple Averaged method and
Distance average method. Simple average method is showing little
higher correlation than Distance weighted average method.
(iv) AOT computed within 100 km radius is having higher correlation than
with lesser radial distances.
(v) AOT computed using 4 neighboring method has less (160) observations
as compared to the method 100 km radius distance (2040).
(vi) TERRA Satellite provide more number of observations than AQUA
satellites
8/10/2019 Air Quality Assessment over Gulf Coast
31/39
TLGVRC Algorithms for spatial estimation of PM2.5 using
MODIS Aerosol Optical Depth
Spatial Estimation of PM2.5
Derive AOT data at PM2.5
observation site using Grid box
method and Radial Distance Method
Establishing Relationship Between
PM2.5and AOT
Estimate Spatial PM2.5using AOT
PM2.5 Observations
MODIS - AOT
Estimated PM2.5 from AOT
8/10/2019 Air Quality Assessment over Gulf Coast
32/39
NASA AURA SATELLITE (launched July 2004)
AuraMLS
TES nadirOMI
HIRDLSDirection of motion
TES limb
Polar orbit; four passive instruments observing same air mass within 14 minutes
OMI: UV/Vis solar backscatterO3, NO2, SO2, HCHO, BrO columns
TES: high spectral resolution thermal IR emissionnadir Ozone, COlimb Ozone, CO, HNO3
MLS: microwave emissionlimb Ozone, CO (upper troposphere)
HIRDLS: high vertical resolution thermal IR emissionOzone in upper troposphere/lower stratosphere
Tropospheric measurement capabilities:
8/10/2019 Air Quality Assessment over Gulf Coast
33/39
33
Researches the composition, chemistry,and dynamics of the Earths atmosphere
as well as study the ozone, air quality,and climate.
HIRDLS: High Resolution Dynamics Limb SounderObserves global distributionof temperature and composition of the upper troposphere, stratosphere, and mesosphere
MLS: Microwave Limb SounderUses microwave emission to measure
stratospheric temperature and upper tropospheric constituents
OMI : Ozone Monitoring InstrumentDistinguishes between aerosol types, such as
smoke, dust, and sulfates. Measure cloud pressure and coverage, which provide data toderive tropospheric ozone.
TES: Tropospheric Emission SpectrometerHigh-resolution infrared-imaging
Fourier transform spectrometer that offers a line-width-limited discrimination of
essentially all radiatively active molecular species in the Earth's lower atmosphere.
Instruments
EOS Aura
8/10/2019 Air Quality Assessment over Gulf Coast
34/39
HIRDLS
multi-channel, microwave radiometer
radiated thermal emissions from the atmospheric limb
spectral intervals in the range (6 to 17) mm, chosen to correspond to
specific gases and atmospheric "windows".
global 3-D fields of atmospheric temperature, several minor
constituents, and geostrophic winds.
34
8/10/2019 Air Quality Assessment over Gulf Coast
35/39
Microwave Limb Sounder (MLS)
The EOS MLS measures thermal emission frombroad spectral bands centered near 118, 190, 240,640 and 2250 GHz
35
8/10/2019 Air Quality Assessment over Gulf Coast
36/39
Microwave Limb Sounder (MLS)
36
8/10/2019 Air Quality Assessment over Gulf Coast
37/39
OMI : Ozone Monitoring Instrument
37
The OMI instrument can distinguish between aerosol types, such assmoke, dust, and sulfates, and measures cloud pressure and coverage,
which provide data to derive tropospheric ozone.
8/10/2019 Air Quality Assessment over Gulf Coast
38/39
TES: Tropospheric Emission Spectrometer
38
Example of TES products
8/10/2019 Air Quality Assessment over Gulf Coast
39/39
Thank YOU
Anjaneyulu. Yerramilli, Ph.D.
Professor & Director
Email: [email protected]
Phone: 601979-3654
Jackson State University
Authors thank the support for Atmospheric
Dispersion Project (ADP) by National Oceanic
and Atmospheric Administration (NOAA) and
NWS through the U.S. Department of
Commerce (SilverSprings,MD); Contract
#NA06OAR4600192