Observations of climate gases from space

1
Observations of climate gases from space

• J.J. Remedios
• EOS-SRC, Dept. of Physics and Astronomy,
  University of Leicester, U.K.
• http//www.leos.le.ac.uk/home
• j.j.remedios_at_le.ac.uk

2
Themes of Lecture

• Greenhouse warming is all about changes in
  radiation balance
• Shortwave (heating) vs longwave (cooling).
• Need to observe the climate gases which affect
  these two terms because
• Changes in concentration lead to surface
  warming/cooling
• For future predictions, need to understand the
  processes which control their concentrations.
• Part I Introduction
• Radiative balance, forcing and uncertainties
• Greenhouse gas trends
• Sources and sinks
• Part II Examples of instruments and observations
• SCIAMACHY CO2
• STRATOSPHERIC O3 AND INFLUENCES
• AIRS H2O

3
First GERB images obtained December 12th 2002
Total radiation
Short-wave only
Illustrate with measurements from the
Geostationary Earth Radiation Budget Experiment
on MSG. The longwave component is given by the
subtraction of the shortwave component from the
total radiation.
4
Radiation balance 5 days of GERB data from
MSG (satellite instrument in geostationary
orbit) Images courtesy of RALNERC/EUMETSAT
TOTAL RADIATION
SHORTWAVE RADIATION
5
RADIATION BALANCE CONCEPT AND CONTRIBUTIONS
6
RADIATIVE FORCING AND SURFACE TEMPERATURE
7
TRENDS IN SURFACE TEMPERATURE I SST rise
(2001/2005) (1850/1899) 0.76C. Last 50
yrs rise 0.13C/decade. Projected is
0.2C/decade From the IPCC reports,
http//www.ipcc.ch
T ANOMALY T - ltTgt
HEATING
IPCC 2001
COOLING
IPCC 2007 updates. Note the separation into land
and ocean Black smoothed mean obsservations Blue
simulations, natural only (solar,
volcanoes) Redsimulations using natural and
anthropogenic
8
TRENDS IN SURFACE TEMPERATURE II From the IPCC
2007 report, http//www.ipcc.ch
Updated radiative forcing diagram from the IPCC
2007 report
9
TRENDS IN SURFACE CO2Ground-based measurements
and ice cores
Average CO2 increase 2 ppmv/yr 0.5 /yr
Courtesy of NOAA CMDL
ICE CORE DATA, IPCC 2007. NOTE THE TIME AXIS OF
10000 YEARS
CO2 is increasing but also CH4 and N2O The rate
of increase is not steady in time but
fluctuates Indicates changing source/sinks in the
biosphere principally
10
Missing aspects The Carbon cycle and feedbacks
for future Earth system understandingWill be
included fully in next generation climate models
of the Earth system. Currently some cycles in
Earth system models of intermediate complexity
(EMICS) but limited spatial resolution
BOXES RESERVOIRS ARROWS FLUXES BLACK
NATURAL PROCESSES, REDANTHROPOGENIC INFLUENCES
Sarmiento and Gruber, Physics Today, 2002
11
GREENHOUSE GAS TREND 7/2001(NOAA CMDL)
CO2 67
CH4 18
N2O 5.9
CFC-11 2.2
CFC-12 5.5
Figure 1. Major long-lived greenhouse gas trends
through the year 2000. The percentages given for
each gas is their estimated respective
contribution to climate forcing by the long-lived
greenhouse gas enhancement since the
pre-industrial period. These five gases account
for about 98.5 of the climate forcing by such
gases. CFC-113 and HCFC-22 (trends not shown)
contribute most of the remaining 1.5.
12
SURFACE AND SATELLITE METHANECH4 larger in NH
than SH. Rice paddy emissions identified as
stronger by satellite data
Courtesy of NOAA CMDL
Average CH4 increase 8 ppmv/yr 0.5 /yr but
has slowed down
MODEL EMISSIONS
SCIAMACHY SWIR CH4
13
Sensitivity of Surface Ts to O3 H2O
O3
H2O
14
Greenhouse gases

• Need to observe the gases which affect longwave
  or shortwave because
• Changes in concentration lead to surface
  warming/cooling
• For future predictions, need to understand the
  processes which control their concentrations.
• Direct anthropogenic forcing
• CO2, CH4, N2O
• CFCs and HCFCs
• SF6, perfluorocarbons
• Indirect anthropogenic forcing
• Natural greenhouse gases O3 and H2O
• Indirect chemical influences CO, NOx,NOy,VOCs.
• In general, want both surface concentrations and
  vertical profiles.

15

• THE PLANCK FUNCTION IS AT THE HEART OF
• OBSERVATIONS FROM SPACE AND CLIMATE CHANGE!
• BUT THE DIRECTION IN WHICH THE
• INSTRUMENT LOOKS IS ALSO IMPORTANT FOR
  OBSERVATIONS!

16
Plancks Radiation Law Earth and Sun
SUN
EARTH
U/V
VIS
TIR
SWIR
FIR
MICROWAVE
17
Plancks Radiation Law I

• One of the classic laws of Physics!
• The Planck function describes the energy flux,
  B(?,T), emitted by a perfectly absorbing/emitting
  (black) body in a small increment of wavelength
  dl
• At all wavelengths, B(?,T), is isotropic, i.e.
  the same in all directions.
• B(?,T) depends only on temperature at a given
  wavelength.
• Two forms (inter-convert using c f l)
• Wavelength (l)
• B(l,T) d l 2hc2 d l
• l5exp(hc / k lT) - 1
• Units of B are W m-2 sr-1 m-1 W m-3 sr -1
• Frequency (f)
• B(f,T) d f 2hf 3 d f
• c2exp(hf/kT) 1
• Units of B are W m-2 sr-1 s-1 W m-2 sr 1 Hz
• B(?,T) d? is not the same as B(f,T) df!

18

SATELLITE ORBITS/OBSERVATIONGEOMETRY

•  

Generalisations NADIR Views downwards Greater
sensitivity to lower troposphere High spatial
resolutuon Low vertical resolution (although 1 km
can be achieved for T, H2O else 5 to 10 km)
Generalisations LIMB Views tangentially Less
sensitivity to middle to lower troposphere Low
spatial resolutuon Very high vertical resolution
(up to 1 km)
NADIR
LIMB
19
OBSERVING CO2 FROM SPACE

• Based primarily on work done by Michael Barkley,
  Alan Hewitt and Paul Monks
• EOS, University of Leicester, Leicester, UK
• But also Michael Buchwitz, University of Bremen.
• And also contributions from many international
  groups
• Papers
• Barkley et al., ACP, 2006 a, b, c ACP, 2007 GRL
  2006
• Buchwitz et al, ACP, 2006, 2007.

20
SURFACE CO2 LOCATIONSIn situ observations
locationsVery good dataWould like satellites
for complete viewAcknowledgement NOAA
Golbalview
21
TRENDS IN SURFACE CO2 ISeasonal cycles
selected surface sites. SH tropics courtesy of
Global Tracker. Red dots are data
South Pole. Little seasonal cycle much less
vegetation Strong trend but with little variation
remote from source (lowest values)
Tropics Seasonal cycle up to 10 ppmv or 2.5
22
TRENDS IN SURFACE CO2 IISeasonal cycles
selected surface sites. NH courtesy of Global
Tracker. Red dots are data
Northern high latitudes. Seasonal cycle of 15
ppmv. Strong summer decrease. Suggests a summer
vegetation sink
Russia Similar seasonal cycle but not as
steep. High values reaching 390 to 400 ppmv.
Possibly an extra source, e.g. fires
23
LATITUDINAL SURFACE CO2Strong seasonal cycle in
NH compared to SHTrend everywhereSource
Globalview
24
SCIAMACHY I
SCIAMACHY LIMB AND NADIR
25
SCIAMACHY SHORTWAVE INFRA-RED OR SWIR
Courtesy of H. Bovensmann University of Bremen
26
Cloud Filter SPICI (SRON) (Krijger et al, ACP,
2005)
A priori Data CO2 profiles taken from 2003
climatology (Remedios, ULeic) ECMWF temperature,
pressure and water vapour profiles A priori
albedo - inferred from SCIAMACHY radiance as a
f(SZA) A priori aerosol (maritime/rural/urban)
SCIAMACHY Spectra, geolocation, viewing geometry,
time
Raw Spectra
Process only if cloud free, forward scan, SZA
75?
SCIATRAN (Courtesy of IUP/IFE Bremen) LBL mode,
HITRAN 2004
Calibration Non-linearity, dark current, ppg
etlaon
I - Calibrated Spectra
I0 Frerick (ESA)
Reference Spectrum weighting functions (CO2,
H2O and temperature)
SCIAMACHY Spectrum (I/I0)
WFM-DOAS fit
CO2 Column (Normalise with ECMWF Surface
Pressure) Accept only Errors lt5, Range340-400
ppmv
Leicester FSI WFM-DOAS Barkley and Monks
27
Example FSI spectral fit

• Process is essentially to fit to spectral lines
  in measured spectra
• Reasonable fit to CO2 lines (top plot)
• Good sensitivity to CO2 concentrations

28
Averaging Kernel
THE AVERAGING KERNELS, A, REFLECT THE TRUE
VERTICAL SENSITIVITY OF THE RETRIEVAL. A dx
(retrieved) / dx (true) It tells us which
heights may have contributed to the data value
that we retrieve. i.e. it represents how much the
retrieved mixing ratio, x, at each retrieval
height, z (retrieved) changes when the true
mixing ratio changes at each true z in turn. If
we apply A to high resolution model atmosphere
profiles then we would obtain the retrieved
concentrations that the satellite would ideally
determine if it flew over the same model profile.
29
Averaging Kernel

• WHY IS A NOT SINGLE VALUED?
• The radiative transfer means that lots of
  altitude levels contribute to the total signal.
• Depends on
• Geometry
• Wavelength i.e. spectroscopy of gases in
  atmosphere
• Concentration of gases
• Effects of aerosol scattering, albedo on changing
  the pathlength or reflectivity respectively
• THE CO2 AK SHOWS NICELY THAT THERE IS SENSITIVITY
  TO THE SURFACE IN THE SWIR
• DIFFERENT SATELLITES HAVE DIFFERENT SENSITIVITIES
  EVEN DIFFERENT RETRIEVALS OF THE SAME PRODUCT

30
Comparisons to ground based FTIR data

• FTIR spectrometer based at Egbert, Canada
• Location 44.23?N , -79.8?E Altitude 251 m
• Accuracy of CO2 columns molec/cm2 ? 8.9
• During 2003
• 74 FTIR measurements
• 5150 valid FSI retrievals
• Large grid 10.0 lon, 2.5 lat of station
• Small grid 5.0 lon, 2.5 lat of station
• Compare FSI columns to 3rd order polynomial fit
  to FTIR data (see Dils et al., ACPD, 2005)
• Normalize FTIR with ECMWF pressure
• Compare to final FSI product

31
Comparisons to ground based FTIR data
32
Comparisons to in situ surface data
Red SCIA (FSI) Blue correlative Green a
priori
33
Comparisons to in situ tower data
Red SCIA (FSI) Blue correlative Green a
priori
34
Comparisons to the TM3 model

• The TM3 model
• Atmospheric transport TM3, driven by NCEP
  meteorological data
• Fossil fuel CO2 emissions
• Ocean air-sea fluxes
• Terrestrial biosphere

• Comparison approach
• Model adjusted for optimal match with in situ
  observations at the South Pole
• i.e. calibrated
• Model is sampled at times locations of
  observations
• SCIA/FSI averaging kernel has been applied to
  model data
• FSI/TM3 retrievals averaged onto 1x1 grid
• Time series of monthly scene averages
• Spatial distribution

TM3 vs SCIA
Canada/Alaska
35
TM3 vs SCIA
Canada/Alaska
Gobi desert
36
Correlation between time series typical greater
than 0.7 (Note No scaling of FSI data) Typical
-2 bias in FSI yearly means -2 difference
(though Gobi Desert ? -1) Bias of TM3 to FTIR
data (using same method) -2 Assuming model
FTIR correct Bias of approx.-4 of FSI CO2 to
true CO2. Relative accuracy to mean approx 1-2
TM3 vs SCIA
Canada/Alaska
Gobi desert
37
(No Transcript)
38
Satellite CO2 I SCIAMACHY SHORTWAVE INFRA-RED
NEAR 2 mm
Strong decreases of CO2 over North America and
Russia in July (summer)
Measure of greenness
39
Satellite CO2 II
Higher CO2 over North America and Russia in
October Note lower values over ocean ocean
sink Note high values over Northern India
Measure of greenness
40
Satellite CO2 III strong vegetation link
(decrease of CO2)
Carbon Tracker CO2 fluxes
41
SCIA vs AIRS Time Series

• CBL ? 396 m TV Tower near Park Fall, Wisconsin
• FT ? Aircraft flights near Carr, Colorado
• (Hurwitz et al, J. Atmos Sci.,2004)

• For North America 2003
• SCIAMACHY
• AIRS

SCIAMACHY Seasonal Cycle resembles CBL i.e. lower
troposphere. AIRS Seasonal Cycle resembles FT
i.e. mid-troposphere
42
N.H. trends of CO2 Buchwitz et al., 2007
43
Trends of CO2 20N, 53N Buchwitz et al., 2007
44
Summary

• Example of climate gas retrieved using scattered
  sunlight in this case in the shortwave
  infra-red (SWIR)
• Sensitivity to CO2 in the lower troposphere.
• Encouraging accuracy results
• but still a long way to go...
• Improve
• A priori data (e.g. use TM3 CO2 profiles ?)
• Calibration
• Aerosols ? (e.g. Houwelling et al, 2005)
• Cloud contamination
• Dual retrieval with Oxygen (O2)
• CO2 ppmv ( CO2 / O2 ) x 0.295
• Can we measure atmospheric CO2 from space?
• Yes !
• Can see seasonal cycles, trends with time,
  regional variations.
• First (tentative!) steps to identify surface
  sources/sinks and to provide modellers with CO2
  satellite data
• Hints of influence of vegetation on CO2
• CO2 retrieved from SWIR is more sensitive to
  lower troposphere

45

• STRATOSPHERIC OZONE AND
• THE OZONE HOLE
• DIRECT RADIATIVE FORCING ANDCLIMATE-CHEMISTRY
  INTERACTIONS
• USE SCATTERED SUNLIGHT IN UV-VISIBLE. MAINLY
  OBTAIN TOTAL COLUMN WHICH IS STRATOSPHERIC
• INTRODUCE IDEA OF VERTICAL PROFILE INFORMATION

46
THE OZONE HOLE IBUV and TOMS measure ozone
column in the U/V
47
GLOBAL OZONE TREND
48
THE OZONE HOLE IIFROM U/V INSTRUMENTS
49
Area of Antarctic Ozone Hole
50

TOMS OZONE LOSS OCTOBER (ANTARCTICA)
MARCH (ARCTIC)

•  

51

TOMS OZONE LOSS 2000 OCTOBER (ANTARCTICA)
MARCH (ARCTIC)

•  

SEPTEMBER 2000
WINTER 1999/2000
52
Ozone Profile

• In situ meaasurements show vertical structures
• Can obtain this from satellites by
• clever nadir sounding (multi-l, high spectral
  resolution, sophisticated retrieval techniques)
  see H2O later
• Combination of scattered sunlight and thermal i/r
  hinted in CO2
• Limb sounding (strat., upper trop.) see
  chemistry lecture for details

53
MIPAS PSC observations ILIMB INFRA-RED
MIPAS
54
Chlorine destruction of ozone SH
MLS O3 October 10 1997
MLS ClO August 27 1997
MLS operates in the LIMB microwave
55
Chlorine destruction of ozone NH MLS
56
VERTICAL PROFILE OF CFC-12
MIPAS LIMB measurements of CFC-12 The transition
from constant mixing ratio to decaying profile is
important The decay rate of the profile is
important
57
AVERAGING KERNELS OF CFC-12
MIPAS limb FTS 0.025 cm-1 spectral resolution
operating in thermal infra-red. CFC-12/Aerosol
joint retrieval. Moore et al., PhD thesis 2005
ASR, 2006

• CFC-12 (a) tropical (b) mid-latitudes (c)
  polar summer (d) polar winter.

58
TRENDS OF STRATOSPHERIC HCFC-22
MIPAS limb FTS 0.025 cm-1 spectral resolution
operating in thermal infra-red. HFC-22/Aerosol
joint retrieval Moore and Remedios, ACP
2008 Since 1994 (ATMOS), both HCFC-22 and N2O
have risen. HCFC-22 has risen a lot!
MIPAS 2003
ATMOS 1994
59
Ozone chronology

• 1974. Rowland and Molina publish paper on CFCs
  and ozone.
• 1978. U.S. bans CFC use in aerosol cans (also
  Canada, Sweden, Norway).
• 1980 EC Council of Ministers vote for freeze on
  production
• of CFCs and 30 reduction in use
  for aerosol cans.
• 1985 Vienna Convention for the Protection of the
  Ozone Layer.
• 1985 First report of the Antarctic ozone hole
  (Farman et al.).
• 1987 Montreal protocol. Production of CFCs cut
  to 80 of
• 1986 levels by 1994 and 50 of 1986 levels by
  1999.
• 1988 Lady Thatcher commitment to ozone layer
  protection.
• 1990 London Amendment --- total phase out of
  CFCs by 2000.
• 1992 Copenhagen Amendment --- Multilateral Fund
  for developing countries established.

60
Chlorine Loading
61
GLOBAL OZONE TREND
SOURCES
STRATOSPHERE
Trend up to 0.4/yr up to 1994 but exaggerated
WMO/UNEP 2006 REPORT ON OZONE LAYER (Fig.1 Exec.
Summary) ODS Ozone Depleting Substances (mainly
chlorine-based) 4 average depletion
1997-2005 Small changes make a big difference
TOTAL OZONE
ULTRA-VIOLET
62
Ozone Policy

• Models and observations have demonstrated clearly
  the effect of CFCs
• Industrial technology development occurred
  pre-protocol
• Precautionary principle resulted in initial
  agreement and successive revision as evidence for
  significant effects strengthened.
• Management by regulation of end games.
• Relatively minor social implications apart from
  impact of consumer pressure.

63

• HEIGHT-RESOLVED WATER VAPOUR IN THE TROPOSPHERE
• DIRECT RADIATIVE FORCING ANDWEATHER FORECASTING

64
NATURAL CLIMATE GASES IN THE TROPOSPHERE

• Can obtain vertical resolution from in the
  troposphere from nadir sounding
• Particularly true for mid infra-red (but also O3
  in UV-visible)
• Many spectral lines (H2O)
• Spectral resolution (O3)
• TES has best spectral resolution
• IASI should do better than AIRS

TES nadir FTS 0.1 cm-1 spectral resolution
operating in thermal infra-red. AK from Kulawik
et al, IEEE Trans. Geos./Remote Sensing, 2006
65
AIRS overview.

• Uses a cross track rotary scan mirror
• 49.5o ground coverage
• 2.67s scan cycle
• 90 ground footprints observed each cycle
• Each 22.4ms footprint contains all 2378 spectral
  samples.
• AIRS IR spatial resolution is 13.5km
• AIRS vis/NIR spatial res is 2.3km
• Global daily coverage, but takes 2-3 days for
  complete global coverage.

66
AIRS

• AIRS is a grating nadir-looking instruments.
• 3 IR wavebands (from 3.74 µm to 15.4 µm) with a
  total of 2378 channels at spectral resolution lt
  1cm -1.
• 4 Vis/NIR channels (from 0.4 µm to 1.1µm).
• IFOV 13.5km at nadir

67
AIRS detector

• Multi-aperture, Echelle grating spectrometer
• Pupil imaging spectrometer.
• 3 IR bands 3.74 - 4.61µm, 6.20 - 8.22 µm, 8.80 -
  15.4 µm (2169-2674 cm-1, 1265-1629 cm-1, 649-1136
  cm-1 respectively).
• Spectral resolution I/DI 1200
• Swath 1650 km
• Spatial res 13.5 km horizontal
• at nadir, 1 km vertical (T, H2O)

68
(No Transcript)
69
So what does AIRS measure?

• Water vapour.
• Surface skin temperature
• Atmospheric temperature
• Greenhouse gasses
• Cloud properties
• Humidity
• Weather.

70
AIRS TOTAL COLUMN WATER VAPOUR
71
AIRS WATER VAPOUR 500 mb to TOA
72
FUTURE SATELLITE ACTIVITY GMES AND RESEARCH
MISSIONS

• What can we do further
• Improve the sensitivity of new CO2 sensors (OCO
  and GOSAT in 2008).
• Improve the vertical resolution/differentiation
  for CO2 shortwave nadir infra-red sensors
  (SCIAMACHY, OCO) with thermal nadir infra-red
  sensors (AIRS/IASI, GOSAT)
• For O3, uv-visible nadir (SCIAMACHY, OMI,
  GOME-2) with TIR nadir (TES, AIRS, IASI) and with
  limb sounders.
• Examine the CO2, CH4 source/sink relationships in
  detail and combine with land/ocean surface
  measurements of vegetation, phytoplankton, SST.
• Inverse model the CO2 and CH4 data to derive
  detailed emission sources
• Long-term monitoring through GMES for trends and
  Kyoto-type protocol verification.
• Assimilate the data to enhance climate gas
  representation

73

• SCIENCE IMPERATIVES FOR CLIMATE GASES CH4 as an
  example
• CH4 SOURCES AND SINKS
• Radiative forcing

Frankenburg et al. Heidelburg, KNMI Science paper

• OBSERVATIONS
• Lower trop SCIAMACHY
• Mid-trop AIRS
• CH4 profiles IASI, IASI/AIRS/SCIA combination,
  limb sounding (MIPAS)