Satellite Remote Sensing of the Troposphere:

1
Satellite Remote Sensing of the Troposphere A
Window on Tropical Ozone Formation
David P. Edwards Atmospheric Chemistry
Division National Center for Atmospheric
Research, Boulder CO, USA
2
Satellite Remote Sensing of the Troposphere

• Satellite tropospheric trace gas measurements
• Combining data from different sensors
• Adding the global perspective linking scales
• Biomass burning Africa and South America
• The tropical tropospheric ozone paradox A
  satellite view
• Future missions

Acknowledgements NCAR MOPITT Team, NCAR/ACD,
Boulder CO, USA Toronto MOPITT Team, University
of Toronto, Canada J. -L. Attie, J. -P. Cammas,
Observatoire Midi Pyrenees, Toulouse, France A.
Richter, University of Bremen, Germany A. M.
Thompson, Y. Ji, Y. Kaufman, NASA GSFC, Greenbelt
MD, USA
3
Satellite Tropospheric Trace Gas Remote
Sensing

• Until now, tropospheric studies have relied on
  field campaigns, regular groundbased and aircraft
  measurements from specific sites, and an
  important input from chemical-transport modeling
• Satellite remote sensing offers the best
  opportunity of making global measurements over
  extended periods of time and will hopefully add
  the larger geographical and seasonal context
• Allows us to address directly two important
  questions
• How do local emissions affect the regional and
  global environment?
• Can we provide a global context to local air
  quality measurements?
• Seeing the troposphere presents major
  challenges that are only just being met
• Current and planned tropospheric sensing
  instruments include Terra/MOPITT MODIS,
  ERS-2/GOME, ENVISAT/SCIAMACHY, Aura/TES OMI,
  OCO
• These instruments provide a new and exciting
  opportunity to test and extend our knowledge of
  global tropospheric chemistry and transport and
  assess whether the new measurements are
  consistent with our current understanding

David Edwards, NCAR
4
MAPS Global CO Measurements

• The MAPS experiment measured large
  mid-troposphere CO levels in the tropics
• This was particularly evident during the October
  1994 flight, at the height of the African burning
  season
• Strong positive correlation was found with
  tropospheric O3

Definition of trends distributions for
tropospheric CO is essential. A satellite-borne
CO sensor operating for extended periods could
help enormously (WMO, 1985)
David Edwards, NCAR
5
Terra/MOPITT

• The Measurement Of Pollution In The Troposphere
  instrument is a thermal and near-IR gas
  correlation radiometer launched on EOS Terra in
  1999
• MOPITT uses a nadir viewing cross-track scan,
  with a 22 x 22 km2 pixel size, and 3 day global
  coverage
• Measures tropospheric CO profile CH4 column

ERS-2/GOME

• The Global Ozone Monitoring Experiment is a
  UV/VIS solar backscatter spectrometer launched on
  ERS-2 in 1995
• GOME is nadir viewing, with a 320 x 40 km2 pixel
  size, and 3 day global coverage
• Measures O3, NO2, BrO, OClO, SO2, HCHO, H2O

David Edwards, NCAR
6
MOPITT MeasurementsCarbon Monoxide

• Created by incomplete combustion or oxidation
  processes including industry, cars, and biomass
  burning
• The main sink of CO is oxidation by OH, which
  usually works as a cleaning agent reacting with
  pollutants to produce soluble species that are
  removed by rain or dry deposition

• Reaction of CO with OH in the presence of NOx
  leads to the formation of tropospheric O3
• CO lifetime is between a week and two months
  depending on location
• This is long enough to be transported without
  becoming evenly mixed so making it a useful tracer

David Edwards, NCAR
Daniel Jacob, University of Harvard
7
The New MOPITT V3 Data 700 hPa 1-12 Nov. 2000
Improved global coverage Improved
calibration Provisional level of validation
The global impact of biomass burning on the
tropical CO distribution is very apparent
David Edwards, NCAR
8
Assimilation of MOPITT CO profile into the
MOZART2 chemical transport model using NCEP
dynamical fields

• TOP MOPITT 700 hPa, Aug. 1-3 2000.
• Elevated CO associated with industrial pollution
  is evident over US China, with significant
  Asian outflow into the Pacific Ocean
• Biomass burning is an important source of CO in
  Africa, South America, and India at this time
• White areas indicate persistent cloud and limits
  to the retrieval
• BOTTOM Sequential assimilation
• into the MOZART2 model
• The high MOPITT measurement density constrains
  the assimilation such that the strength and
  locations of the CO emissions follow closely
  the data

David Edwards, NCAR
9
CO Plumes Observed During TRACE-P

• The NASA/GTE TRACE-P aircraft campaign was
  conducted over the western Pacific during
  Feb-Apr, 2001
• Goal to study the outflow of Asian pollution
  resulting from biomass burning and industry
• MOPITT data over the western Pacific were
  provided to TRACE-P in near-real-time for use in
  flight planning

David Edwards, NCAR
10
Averaging Kernels andVertical resolution

• MOPITT profiles have been compared with aircraft
  in situ measurements taken at CMDL sites and
  during major field campaigns such as SAFARI-2K
  and TRACE-P
• Applying the MOPITT averaging kernel removes fine
  scale vertical structure
• Agreement is generally very good for latest V3
  data version

MOPITT In-Situ Avg Kernel x In-Situ
TRACE-P Flt. 11 2001-03-17
TRACE-P Flt. 16 2001-03-30
David Edwards, NCAR
In situ data from G. Sachse, NASA Langley
11
GOME Tropospheric Vertical Column NO2

• GOME measures total column NO2 and provides an
  estimate of the tropospheric column variation
• The tropospheric retrieval assumes a
  longitudinally homogeneous stratospheric column
  for a given solar zenith angle and a fairly clean
  remote maritime troposphere

• Observed slant column densities are converted to
  vertical column densities using an air-mass
  factor
• This depends on the vertical NO2 profile, aerosol
  loading, cloud cover, and scattering albedo
• The monthly averaged product is most reliable due
  to increased measurement density and improved SNR

David Edwards, NCAR
Andreas Richter, University of Bremen
12
Global HCHO from GOME July 1996
Paul Palmer, Harvard

• GOME data have recently been used to retrieve
  formaldehyde
• HCHO is a principal intermediate in the oxidation
  of CH4 and VOCs in the troposphere and can be
  used as a tracer for hydrocarbon emissions
• CH4, VOCs ? ? HCHO ? ? CO HO2

David Edwards, NCAR
13
GOME and SCIAMACHY SO2

• Mount Etna eruption
• Eruption started on 10/26/02
• Large SO2-plume is seen moving from Sicily
  towards Africa
• On average, GOME (320 x 40 km2) and SCIAMACHY (60
  x 30 km2) agree well (30 min. time difference)
• Maximum values seen by SCIAMACHY are up to a
  factor of 4 larger and plume structure much
  better resolved

Andreas Richter, University of Bremen
David Edwards, NCAR
14
Terra/MODIS Aerosol

• MODIS can distinguish between coarse and fine
  particles using multiple channels from the
  visible to the near IR
• Course aerosol mode Mainly dust
  and sea salt
• Reflects in the VIS NIR Appears red in the VIS
  due to absorption in the blue
• Surface characterization is an issue over land
• Fine aerosol mode Anthropogenic pollution
  and biomass burning
• Reflects mainly in the VIS

Yoram Kaufman, 2002
15
MODIS Aerosol
Distinguishing between coarse and fine mode
particles, Sept. 2000
Coarse mode
(AOT)
Fine Mode
Yoram Kaufman, NASA GSFC
16
MODIS
MODIS Aerosol MOPITT CO

• Reasonable correlation is obtained between MODIS
  fine mode aerosol and MOPITT CO column averaged
  over Sept. 2000
• Both CO and fine mode aerosol are produced by
  urban pollution, industrial combustion, and
  biomass burning

Fine Mode AOT
Yoram Kaufman, NASA GSFC
MOPITT
CO Column
David Edwards, NCAR
17
(No Transcript)
18
Rocky Mountain Fires Terra/MODIS June 10 2002
Hayman
Missionary Ridge
Ponil
David Edwards, NCAR
19
MOPITT 700 hPa CO From Western Fires
David Edwards, NCAR
20
Smoke over Eastern Canada/USA
MODIS fires and smoke
MOPITT CO Column
July 1-8 2002
July 8 2002
Charles Ichoku, NASA GSFC
David Edwards, NCAR
21
Tropospheric Chemistry and Biomass Burning

• Approximately half-billion hectares of natural
  human-induced biomass burning occurs each year
• In tropical regions, biomass burning for
  cultivation, deforestation and savanna grazing is
  a major forcing mechanism for tropospheric
  photochemistry
• Large amounts of CO and CH4 are emitted by the
  smouldering fires and reaction with OH in the
  presence of NOx, either from burning or from
  lightning sources, leads to high levels of O3

• The study of tropospheric ozone is an active
  field due to the key role that this trace gas
  plays in determining the oxidizing capacity of
  the atmosphere
• The investigation of burning emissions and the
  resulting transport and chemistry has been the
  motivation for a number of field campaigns over
  the last decade, most recently SAFARI-2000 in
  southern Africa

David Edwards, NCAR
22
TRMM/VIRS Fire Observations

• Fire pixels are identified according to radiance
  threshold and difference criteria for channels at
  3.75 mm and 10.8 mm
• Because of instrument sampling and cloud
  coverage, it is most reliable to view composite
  maps of several days
• The location of the burning region is seen to
  move South during the year
• A high degree of inter-annual variability is
  observed

Jan. 2000
Tropical Rainfall Measuring Mission Visible and
Infrared Scanner (TRMM/VIRS) 4.4 km2 fire product
Oct. 2000
David Edwards, NCAR
Yimen Ji, NASA GSFC
23
MOPITT CO Mixing Ratio at 700 mb and Fire
Locations, March-September 2000
24
Providing global context to local measurements
Long-range transport of biomass burning products
October 1-15, 2001

• MOPITT and ground-based FTIR CO total column
  measurements taken at Lauder, New Zealand.
• The plume from biomass burning causes a peak in
  the measured CO in the usually clean air over New
  Zealand

FTIR data N. Pougatchev, NASA N. Jones, NIWA
David Edwards, NCAR
25
Combining sensor data to provide a clearer
picture of tropospheric chemistry Tropospheric
Ozone in the South Atlantic

• Analysis of the fire count data from TRMM/VIRS
  shows that the maximum southern burning season
  occurs in ASO
• High EP/TOMS tropical tropospheric ozone values
  correlate well with the burning plume as
  indicated by the high CO levels measured by MOPITT

David Edwards, NCAR
26
The Tropical Tropospheric Ozone Paradox
EP/TOMS tropospheric ozone column from the
modified residual method
(TRMM/VIRS) fire product
Most of the NH biomass burning occurs north of
the ITCZ, while the maximum tropical tropospheric
O3 columns are noted south of the ITCZ leading to
the ozone reversal - Thompson et al., GRL 27,
3317 2000
David Edwards, NCAR
27
What do the Models Show?
MOZART-2 tropospheric O3 column below 133 hPa
Jan. 2001
MOZART-2 shows two ozone maxima, one in the NH
lower troposphere near the burning region and a
SH mid troposphere maximum further south
David Edwards, NCAR
28
MOZART-2 Ozone Jan. 2001
435 mb
700 mb

• Mozart-2 CTM simulation using NCEP winds shows
  that the Atlantic ozone distribution in the early
  part of the year has two maxima
• the first is in the lower troposphere over the
  northern biomass burning region
• the second is due to lightning NOX over Africa
  and South America which forms a mid-troposphere
  O3 plume extending out over the southern tropical
  Atlantic

David Edwards, NCAR
29
MOZART-2 Experiment Biomass Burning Emission
Perturbation
D O3 700 hPa
D O3 435 hPa
700 mb
The effect on the ozone distributions of a 10
increase in all biomass burning emissions over
Africa during DJF 2001 Ozone enhancement is
greatest at low altitude near the burning
region Upper altitude enhancement is smaller and
further South
David Edwards, NCAR
30
MOZART-2 Experiment Lightning Emission
Perturbation
D O3 700 hPa
D O3 435 hPa
The effect on the ozone distributions of a 10
increase in lightning NOx emissions over Africa
during DJF 2001 Ozone enhancement is greatest at
the higher altitude over the Atlantic
David Edwards, NCAR
31
Terra/MOPITT CO
350 hPa

• Very good correlation is seen between the MOPITT
  CO plume retrievals and the positions of the
  springtime fires
• Clear indication of significant convection and
  net southern transport from the burning region
  across the equator and into the Gulf of Guinea

700 hPa
ppbv
MOPITT monthly mean gridded CO for Jan. 01
David Edwards, NCAR
32
Seven day trajectories calculated using NCEP
winds for Jan. 15-22 2001. Positions of
trajectory initialization defined by locations of
the TRMM fire observations with an initial
injection altitude of 3 km
Transport from the Biomass Burning Region
David Edwards, NCAR
33
MOPITT Measurements Indicating Convection

• Low altitude CO is high in the NH near the
  intense burning in East Africa
• Emissions are carried by the Harmattan to the
  ITCZ with subsequent convection and
  interhemispheric transport
• The latitudinal gradient of CO reverses at high
  altitude, with maximum values observed in the SH

CONCLUSION 1 The NH Springtime fire plume does
not extend far enough South to explain the
observed TOMS TTO maximum in the southern
tropical Atlantic
Latitude/altitude CO cross-section at 12 E, Jan
20-27 01
David Edwards, NCAR
34
MOZAIC and MOPITT
MOZAIC flight, 26 Feb., 2001
MOZAIC flight Brazzaville (4.37 S, 15.46 E) to
Douala (4.01 N, 9.72 E). Ozone distribution
along flight path correlates well with MOPITT CO .
David Edwards, NCAR
35
ERS-2/GOME TroposphericVertical Column NO2

• GOME residual tropospheric NO2 vertical columns
  also show good correlation with fire locations
• Resulting distribution is similar to the CO
  plumes observed by MOPITT
• Industrial hotspots are also evident over Lagos
  and Johannesburg
• High NO2 over southern Africa is most likely a
  signature of lightning in the area

Jan. 2001 mean GOME tropospheric NO2
David Edwards, NCAR
36
The Role of Lightning

• Observations of lightning flashes by the TRMM/LIS
  instrument show significant activity over
  southern Africa
• Upper troposphere lightning NOx is a precursor
  to O3
• Ozone is formed over Africa and in the plume
  extending westward into the Atlantic

LIS Lightning Distribution Jan. 2001
CONCLUSION 2 The lightning NOx is most likely
responsible for the tropical SH tropospheric
ozone maximum observed by TOMS
David Edwards, NCAR
37
The Orbital Carbon Observatory

• ISSUES Carbon dioxide (CO2) is the principal
  atmospheric component of the global carbon cycle
• Primary anthropogenic driver of climate change
• Only half of CO2 produced by human activities
  over the past 30 years has remained in the
  atmosphere
• Where are the sinks?
• Will this continue?

Atmosphere

• INSTRUMENT OCO is a 3 channel spectrometer
  measuring CO2 at 1.58 and 2.06 mm with an O2
  0.76 mm channel to derive surface pressure
• GOAL First CO2 column measurements from space
  with a 1 ppmv precision on 5 deg scales at
  16-day intervals

?
?
Human Activity
Land
Ocean
Slide info David Crisp, OCO PI, JPL
38
David Edwards, NCAR
39
Summary

• The new tropospheric satellite sensor data will
  play an increasingly important role in explaining
  chemistry and transport processes
• This will be complimentary to continued in-situ
  measurements and modeling studies
• The potential for combining measurements from
  several sensors provides a powerful tool for
  investigating the tropospheric production of
  ozone precursors
• MODIS offers possibility of combining CO and
  fine/coarse particle measurements to examine the
  chemical origin and transport of anthropogenic
  aerosol
• MOPITT CO provides a clear picture of the
  pollution plumes that result from fires biomass
  burning and can be used to study convection and
  advection processes
• Ongoing studies are using data from several
  satellite sensors to help distinguish biomass
  burning, lightning, and biogenic sources of O3
  precursors to further explain the seasonal
  variation of the tropospheric O3 distribution

David Edwards, NCAR