NESDIS CCNY Collaboration on Air Quality Applications of NOAA Operational Satellite Data Shobha Kond

1
NESDIS CCNY Collaboration on Air Quality
Applications of NOAA Operational Satellite Data
Shobha Kondragunta, NOAA/NESDIS/STARand Fred
Moshray, CCNY
CREST Annual Symposium February 21,
2008 Mayaguez, Puerto Rico
2
NOAA Air Quality Program Structure
Active collaboration with EPA for over 50 years
3
NESDIS Air Quality Program Objectives

• Support NOAA-EPA MOU and MOA which includes the
  development and deployment of operational air
  quality forecast guidance
• Development of algorithms to derive trace gas and
  aerosol products from NOAA operational satellite
  sensors
• Research (NASA) to Operations (NOAA)
• Conduct air quality application studies to
  demonstrate the usability of satellite data in
  air quality applications
• Data analysis and validation
• Modeling and assimilation studies
• Support NWS in air quality forecast verification
  and improvements
• Develop and deploy decision support systems
• Algorithm/product development from future
  satellite sensors
• Mission planning activities
• Building collaborations across multiple
  government agencies and academic institutions

4
Air Quality and Human Health

• Aerosols (PM2.5) and ozone
• Induce respiratory diseases and cancer
• Aggravated asthma
• Chronic bronchitis
• Premature death
• Decreased lung function
• Impacts of Poor Air Quality on Society (in U.S.)
• 60,000 Death per annual (mean)
• 143 Billion Cost per annual (mean)

Both coarse particles and fine particles
penetrate lower regions of the lungs and cause
damage
Science 289, 2000 American Lung Assoc. 2001
5
Air Quality Non-attainment Areas
Ozone

• Noting that several areas in the northeast are in
  non-attainment of EPA standards for both ozone
  and PM2.5, NESDIS partnered with CCNY to explore
  the use of satellite data and in situ
  measurements in studying air quality in this
  region
• CCNY work with NESDIS/STAR by leading and
  supporting
• Interactions with NYDEC
• Validating and improving satellite retrievals
• Applying satellite data with a more regional
  focus

PM2.5
6
LIDAR Observations
UMBC CCNY Hampton UAH Puerto Rico

• Integrating data from models and satellites, to
  scale satellite-observed AOD, requires a thorough
  evaluation of model vertical profile
• Data availability (especially vertical profile)
  for model evaluation continues to be a challenge
• Currently, only few LIDAR sites operate on a
  semi-routine basis in the U.S.
• Potential CALIPSO application

Northeast
Southeast
Tropical
7
Air Quality Products

• Aerosol Optical Depth
• Proxy for PM2.5
• Trace gases (NO2, HCHO, CHO-CHO, SO2, BrO, CO)
• Precursors to ozone and some are EPA criteria
  pollutants
• Biomass burning emissions
• Episodic events that inject tons of PM2.5 into
  the atmosphere and lead to violations of EPA
  standards in the vicinity and downwind of fires.
  Useful for forecasting applications as well as
  monitoring EPA exceptional rule

Impact of forest fires on CMAQ aerosol predictions
8
Plume Detection Optical Properties
9
Smoke Plumesfrom fires transportedover USAug
13-15 2007from UV aerosol Index from OMI
10
GOES , AERONET AOD and Aloft Plume Layer
Interaction
Aug.15, 2007 (Plume/PBL Interaction)
Aug.14, 2007 (Aloft plumes)
Aug.13, 2007 (clear)
11
Onset of Plumes actually begin before high PM2.5
on 08-14-07 Column Properties
Same time unit Local timeEST
12
AERONET Column Angstrom Coefficient
13
Separating PBL and Plume AODPlume AOD gtgt PBL AOD

Good match with AERONET SP measurements
Fraction of plume layer AOD in the total column
PBL AOD .3 at 1400 hrs
14
CALIPSO Lidar for Plume Tracking

• Application Validation

15
Smoke-plume
Date2007-08-15, Time0707 UTC Smoke-plume
event from Idaho/Montana
16
Comparison of aerosol-plume extinction profiles
between CALIPSO-CCNY measurement
CALIPSO analysis at 532-nm Plume_AOD0.588 (/-
0.087) Total_AOD0.642 Lidar ratio S170 sr
(plume)
17
Plume optical depth at 532-nm
Depolarization ratio in plume layer
18
Plume Mixing with the Boundary Layer

• Impact on Air Quality
• (AOD and PM2.5)

19
Plume spatial-temporal distribution measured by
CCNY-Lidar
Lidar returns at 1064-nm
Plume Mixing Into PBL
20
Particulate Matter
Observation of low PM2.5, unlike satellite data
would imply
PM estimator Eliminating plume
21
MOPITT CO Profile
22
Improvement of Satellite AOD Retrievals
23
2007, July 4
surface reflectance
WEST
EAST
1315
1915
west
2215
east
24
MODIS and Aeronet AOD Comparison (10km) using C005
Significant overestimate seen in operational
product
25
Operational MODIS Reflection Ratios
Urban Albedo Ratio is some what Larger than old
Collect 4 but smaller than what we see In
Hyperion for urban pixels (.75-.80) Vegetation
albedo Is more stable around collect 4 value
of.5
26
Ground Reflectance ratio (VIS 670nm/MIR
2120nm)Derived by RT using sunphotometer optical
depths composited (02-07) from Terra MODIS for
fine mode dominated days with stable AOD.
Studies show that the ratio is not angle
dependent within the measurement error
When exploring the correlation, we clearly see
the Urban hot spot as well as very large values
over water (since water absorbs at 2130) Also,
correlations are closer to C005 to the north in
vegetated areas
27
AOD Retrieval with refined model at different
resolutions
Masked water
C005 10km
Urban 10km
MODIS
Aeronet
Aeronet
Urban 3.0 km
Urban 1.5 km
Aeronet
Aeronet
28
10-03-2006 AOD retrieval at 1.5 km resolution
Significant Removal of Hot Spot AOD. Note also
we have performed retrieval over water using
Water algorithm and reasonable water leaving
reflectance Showing seemless transition over water
29
SO4 dominated air masses have much higher AOD but
only a bit higher PM compared to NO3 (NO3 mainly
winter so lower total aerosol load for given
surface PM conc.) OC greatest fraction as
coarse. Fine mode peak shifted slightly to
larger particles for SO4 compared to NO3, OC.
Analysis has implications for satellite
retrievals (aerosol model assumptions in the
algorithm). Similar analysis with CCNY data will
provide perspective for eastern US
30
Original GOES AOD image transformed to smoke AOD
image
Smoke AOD converted to smoke concentrations
Original GOES AOD Image with fire hot spots in
red circles. Clouds are shown in white. Gray
area is where there is no AOD retrieval

• GOES smoke concentration product tracks
  long-range transport of smoke into the U.S. CCNY
  LIDAR and other observations can be used to (1)
  identify if smoke detection is correct, (2)
  altitude at which smoke is being transported, (3)
  if smoke is impacting surface air quality

31
GOES AOD Assimilation Experiments

• July 30 to August 4, 2007 regional scale sulfate
  episode was simulated using Community Multiscale
  Air Quality (CMAQ) model
• Assimilation of hourly GOES AODs to tune CMAQ
  PM2.5 initial conditions shows improved
  correlation between predicted and observed
  surface PM2.5 concentrations
• Assimilating hourly information helps tune the
  model for missing sources and sinks and improves
  predictions

32
NO2B
NO2C
GOME-2 Tropospheric Column NO2
GOME-2 Vertical Column NO2
GOME-2 Slant Column NO2
NOAA GOME-2/OMI NO2 Work
Goal To process GOME-2 and OMI data with a
common algorithm to study diurnal variations in
tropospheric NO2 over CONUS Common Algorithm
Harvard SAO GOME slant column NO2 algorithm
modified to run on both GOME-2 and OMI. NASA
GSFC OMI NO2B and NO2C algorithms to convert
slant column NO2 to vertical column density and
removing stratospheric NO2 from total column to
obtain tropospheric NO2 amount Progress (1) Test
processing of the common algorithm on GOME-2 data
for August 2007 complete, (2) Test processing of
the common algorithm on OMI data for August 2007
underway Ongoing and future work (1) Optimize
the algorithm (e.g., surface reflectivity
database, NO2 cross sections, a priori profiles),
(2) conduct spatio-temporal analysis and
verification of GOME-2 and OMI NO2 retrievals for
summer 2007, (3) compare GOME-2 and OMI NO2
retrievals with NWS operational CMAQ NO2
predictions Other applications Work with state
environmental agencies (e.g., NYDEC) in using OMI
and GOME-2 NO2 products in SIP (State
Implementation Planning) modeling
33
NWS Global Aerosol Forecasting Efforts
Ho-Chun and McQueen, personal communication
1 The aerosol modeling will be added later
34
Future Directions

• BRDF for GOES Imager Visible Channel
• AQUA MODOIS Ground Reflectivity Maps, Comparison
  to TERRA results
• GOES-MODIS Data fusion
• PM2.5 Estimator Dependence on Aerosol type and
  Meteorological Parameters
• Trace Gas-Aerosol Correlations
• High Resolution Modeling in New York Metro Region

35
470/2120 ratio