Ground-based Measurements

1
Ground-based Measurements

• Measurements
• Retrieving the Desired Information
• Comparison Between Instruments
• Satellite Validation
• Toward Model-Measurement Comparison

Prepared by Dr. Stella M L Melo University of
Toronto
Part of the material presented here was provided
by Prof. K. Strong
2
Points

• Understanding what is measured
• Extracting the desired information from the
  measurement
• Comparing measurements made using different
  techniques
• Measurements and models

3
Measurement

• Remote sensing/sounding
• a means of obtaining information about an object
  or medium without coming into physical contact
  with it.
• This generally involves the measurement of EM
  radiation that has interacted with the object or
  medium of interest.

4
Interaction of Radiation with Gases

• Ionization dissociation (UV)
• Electronic transition (UV- visible)
• Vibrational transitions (infrared)
• Rotational transitions

5
Measurement

• Remote sensing/sounding techniques can be
  classified by the type of measurement and the
  spectral region.
• ACTIVE vs. PASSIVE
• IMAGING vs. NON-IMAGING vs.
• WAVELENGTH OF RADIATION MEASURED

6
ACTIVE
RADAR
(m-WAVE)
FABRY-
FOURIER
GRATING
PEROT
TRANSFORM
(UV/VIS/IR)
(VIS/IR)
(VIS/IR)
7
Measurement

• Passive Systems
• These generally measure the extinction, emission,
  or scattering of radiation in order to retrieve
  atmospheric properties.

1. Radiometers isolate bands of natural radiation
   using some form of spectral filter
2. Spectrometers disperse natural radiation into
   its constituent wavelengths over a finite
   spectral range or use interference effects to
   obtain spectral information
3. Monochromatic LASER Techniques - LASER is used
   in a passive mode in a heterodyne system, serving
   as a local oscillator

8
Measurement

• Active Techniques
• (1)- radio (microwave) spectrum ? RADAR (RAdio
  Detection And Ranging)
• (2)-visible and infrared spectrum
• LASER (Light Amplification by Stimulated Emission
  of Radiation)
• LIDAR (LIght Detection And Ranging)

9
Gas Absorption/Emission Spectroscopy

• Absorption and emission spectra provide a means
  of identifying and measuring the composition and
  temperature of the atmosphere.

http//ess.geology.ufl.edu/ess/notes/050-Energy_Bu
dget/absorbspec.jpeg
10
Gas Absorption/Emission Spectroscopy
11
Gas Absorption/Emission Spectroscopy
Four processes will change the intensity of the
EM radiation as it passes through the volume A
absorption from the beam (depletion term) B
emission by the material (source term) C
scattering out of the beam (depletion term) D
scattering into the beam (source term)
12
What is measured
Passive remote sounding instruments are used to
observe extinction, emission, and scattering of
natural radiation.
Scattering
Extinction
Emission
13
Measurements Techniques

• DOAS
• Retrieving vertical distribution of constituents
  from ground-based measurements
• Temperature form Airglow and from Lidar

14
Ground based Measurements DOAS

• Differential Optical Absorption Spectroscopy
  DOAS
• a method to determine concentrations of trace
  gases by measuring their specific narrow band
  absorption structures in the UV and visible
  spectral region Platt and Perner, 1983 Platt,
  1994.

15
DOAS

• Differential absorption cross-sections of some
  trace gases absorbing in the UV/vis wavelength
  region.
• On the right axis the detection limits of the
  trace gases is listed together with the typical
  light path lengths used to measure them.

16
DOAS
Varies slowly with l
17
DOAS NO3

• Example of NO3 analysis
• Daytime reference and night time spectra to be
  analyzed
• The division of the two spectra with a low-pass
  filter
• The fit of NO3 to the OD spectrum, d) Residual
• e) and f) NO3 and H2O cross-section

From Allan et al., JGR 105, 2000
18
Viewing Geometry

• Direct sun
• Zenith sky
• Bright-sky (less commonly used?)

Zenith Sky
Direct Sun
19
Viewing Geometry

• Direct Sun
• Longer path through the troposphere (than zenith)
• Air mass factor calculation normally requires
  only geometric considerations
• Clear sky conditions are required
• Zenith Sky
• Longer path through the stratosphere
• Possible to measure absorption at SZA greater
  than 90
• Any weather conditions
• Air mass calculation involves complex scattering
  geometry
• Bright-Sky
• Instrument points towards the brightest part of
  the sky
• Increased sensitivity to aerosol scattering

20
Zenith sky configuration

• Particularly sensitive to stratospheric absorbers
• For the analysis of DOAS measurements using
  direct or scattered solar radiation traversing
  the atmosphere, the so-called air mass factor
  concept has been developed (see e.g. Noxon et al.
  1979, Solomon et al. 1987b).
• Since the measured slant column density (SCD),
  the integrated trace gas concentration along the
  light path, strongly depends on the solar zenith
  angle (SZA), it is advantageous to convert the
  SCD into a vertical column density (VCD), the
  vertically integrated trace gas concentration.
• This conversion is usually performed by dividing
  the SCD by the air-mass factor (AMF)  
  
• VCD SCD(SZA) / AMF(SZA)
  (apparent SCD)

21
DOAS NO3
Figure 3. Examples of the variation in the
vertical column density of NO3 obtained from
zenith sky measurements during field campaigns
carried out at Weybourne, Mace Head, Tenerife,
Cape Grim and Norway.
From Allan et al., JGR 107, 2002
22
DOAS NO3
Figure 6. Examples of profile retrieval from NO3
vertical column density measurements made on 3
February 1999 (triangles) and 15 February 1999
(circles) at Cape Grim, Tasmania.
From Allan et al., JGR 107, 2002
23
U of T spectrometer Zenith Sky
24
O3 DOAS fit
By Elham Farahani
25
Ozone Eureka 1999
Fig. 1. Ozone vertical column densities measured
with the University of Toronto UV-visible
spectrometer, compared with data from sondes, a
Brewer spectrophotometer operated by the
Meteorological Service of Canada, and the TOMS
satellite instrument.
Melo et al., ASR, 2004
26
NO2 DOAS fit
By Elham Farahani
27
NO2 Vertical column
Fig. 2. NO2 vertical column densities at SZA90
observed with the UV-visible spectrometer. Values
were obtained assuming a reference column density
of 1.46x1016 molec.cm-2.
Melo et al., ASR, 2004
28
NO2 Vertical distribution
NO2 slant column (x1017 molec cm-2)
(K. E. Preston et al., J. Geophys. Res. 102,
19089, 1997)
29
Retrieval or inversion theory

• The direct or forward problem
• - A detector measures a signal S f(T) which is
  generated by the interaction of radiation with
  the target (atmosphere, clouds, etc.).
• ? given properties of the target, calculate
  signal
• The inverse problem
• Want to determine properties of the target,
  given by the inverse function T f-1(S).
• ? given signal, calculate properties of the target

30
Solving the inverse problem

• Solving the inverse problem is complicated by a
  number of difficulties
• (1) Non-uniqueness of the solution
• several unknown parameters, which can be combined
  in different ways to generate the same observed
  signal. i.e., have several solutions T1 f-1(S),
  T2 f-1(S), etc.
• (2) Discreteness of the measurements when the
  measured quantity is a smoothly varying function.
  e.g., T is a function of height z, while S is
  measured at discrete levels over some range of
  heights, so where K(z) is called a kernel
  function or a weighting function.
• (3) Instability of the solution due to errors in
  the observations S. e.g., If ? is the error on S,
  then where ? can produce a large change in the
  retrieval of T(z).

31
NO2 Vertical distribution

• The inversion problem consist of expressing the
  unknown profile x in terms of the observations y

y measurements - SCD F forward model b
model parameter ey measurements error
yF(x,b)ey
Direct inversion? Well, the problem is
under-constrained
There is no unique solution choose the
optimal solution
32
NO2 Vertical distribution
- For the purpose of retrieval, the inverse
problem can be reformulated in a discrete form
(Rodgers, 1976)
yKx where Kdy/dx
The rows of the K matrix are know as weighting
functions and each one corresponds to a different
measurement.

• Rodgers (1976) reviews a number of different
  methods
• Sequential estimation was chosen for this study
• Solving the optimal estimation equation
  sequentially.

33
NO2 Vertical distribution

• Vertical profile can be retrieved using optimal
  estimation sequentially Rogers, 1990
• minimizes the difference between measured and
  calculated slant column values as a function of
  SZA to determine the optimum solution profile for
  each set of twilight spectra.
• is based on sequentially combining a set of
  independent measurements by taking a weighted
  average, using the measurement errors as weights.
  
• does not require iterations
• includes a formal treatment of errors.
• Vertical resolution 5 to 7 km from 10 to 35 km
  altitude.

34
Retrieval

• Optimal estimation
• Combining two independent measurements, x1 and
  x2, by taking the average, using the inverse of
  the error covariances, S1 and S2, as weight

Covariance matrix of x
- Those equations can then be used to invert
slant column measurements
35
Retrieval
xK-1y The measurement error inverse covariance
S-1e can be mapped into profile space using the K
matrix KTS-1eK. x0 a priori with error
covariance Sx


Covariance of x
Simplifying
Can be solved sequentially
36
Retrieval

• The measurements are treated as m scalars, yi,
  with corresponding standard deviation si.
• The a priori, x0, is the first guess of the
  vertical profile
• It is sequentially updated or improved using one
  weight function Ki and one measurement yi at a
  time
• The error covariance S is retrieved
  simultaneously with the vertical profile

37
Does it works?

• CMAM vertical profiles of NO2 concentration for
  Vanscoy for a set of SZA range during both
  twilights are used to generate SCD as a function
  of SZA (Chris McLinden).
• Those SCD are used into a retrieval code to
  retrieve back the NO2 vertical distribution at
  sunrise and sunset (SZA90)
• Both the forward model and the box model employed
  in the retrieval are independent from either CMAM
  or Chris McLinden models.
• The a priori used in the retrieval is also
  independent of CMAM.

38
Testing NO2 retrieval
39
Averaging Kernels
40
Information content

• Smoothing CMAM using Rodgers aproach (Rodgers and
  Connar, JGR 108, Vol D3, 4116, 2003) assume CMAM
  is the true (Xh), the smoothed profile (Xs)
  will be given by
• XsXa A(Xh Xa)
• Xa represents the a priori adopted in the
  retrieval and A represents the Averaging Kernels.

41
Testing NO2 retrieval
42
Adding Tropospheric NO2
Zenith viewing geometry low sensitivity to
troposphere
43
Using measurements
Fig. 4. NO2 a priori, measured, and retrieved
absolute slant column densities for March 31,
1999 (day 90) measured at Eureka. Values were
obtained assuming RCD1.46x1016 molec.cm-2.
Melo et al., ASR, 2004
44
Using real measurements
Eureka 1999
Fig. 5. Left plot retrieved NO2 profiles for
SZA90 on March 31st, 1999 (day 90) compared
with the initial profile. Right plot error in
the a priori and retrieved profiles. Values were
obtained assuming RCD 1.46x1016 molec.cm-2.
Melo et al., ASR, 2004
45
NO2 and O3 Eureka 1999
Fig. 7. O3 partial pressure measured using a
sonde launched from Eureka at 2327 UT on March
31, 1999 (day 90).
NO2 Vertical distribution
46
NO2 from different platforms
Figure 1. The NO2 slant columns measured with the
ground-based UV-visible spectrometer on August
24, 1998 (MANTRA) and the fitted values. The
symbols represent measured values while the lines
represent retrieved values. Lower plot the
percent difference ((measured-fitted)/measured),
sunrise (diamonds) and sunset (squares).
Melo et al., A-O, Submitted
47
MANTRA Balloon Campaign

• MANTRA is conducted at Vanscoy, Saskatchewan
  (52N, 107W)
• - August 24, 1998,
• - August 29, 2000,
• - September 3, 2002,
• - NEXT August 2004

• Instruments of interest here balloon-based SPS,
  SAOZ, and DU-FTS, the ground-based UT
  Spectrometer and the sondes.
• In this work we report on vertical profiles of T,
  O3, NO2, N2O, CH4, and HCl concentrations .

48
MANTRA Scientific Objectives

1. To measure vertical profiles of the key
   stratospheric gases that control the mid-latitude
   ozone budget.
2. To combine these measurements with historical
   data to quantify changes in the chemical balance
   of the stratosphere, with a focus on nitrogen
   compounds.
3. To compare multiple measurements of the same
   trace gases made by different instruments.
4. To use the measurements for validation and
   ground-truthing of Odin, ENVISAT, and SCISAT-1.

http//www.atmosp.physics.utoronto.ca/MANTRA
49
MANTRA 1998
50
NO2 - Comparing with models

• What kind of models?
• CTM - MEZON (Model for Evaluating oZONe Trends)
  global chemistry-transport model.
• horizontal resolution of 4 in latitude and 5 in
  longitude
• Vertical the model spans the atmosphere from the
  ground to 1 hPa
• driven by wind and temperature fields provided by
  the UKMO
• initial conditions for the trace-gas
  concentrations have been taken from the UARS
• lightning source of NOx is prescribed
• Rozanov et al. (1999) and Egorova et al. (2001,
  2003)
• GCM CMAM Canadian Middle Atmosphere Model
  (CMAM)
• upward extension of the Canadian Centre for
  Climate Modeling and Analysis (CCCma) spectral
  General Circulation Model (GCM) up to 0.0006hPa
  (roughly 100 km altitude).

51
NO2 diurnal variation - CMAM
52
NO2 MANTRA MEZON (CTM)
53
NO2 - Comparing with GCM
- NSEP/NCAR (1970-2001), UKMO (1993-2002) and
CMAM (24 years) long-term means of zonal wind
velocity over Vanscoy. (Wunch et al, A-O in press)
54
NO2 MANTRA - CMAM
55
NO2 - Comparing with CMAM
MANTRA Mid-latitude summer time
56
NO2 - Comparing with CMAM
MANTRA Mid-latitude summer time
57
MANTRA CMAM
1998 Campaign
58
Reaction pathway between the principal NOy
constituents
Gao et al, GRL, 1999.
59
Nitrogen Partitioning MANTRA - CMAM
- Dynamics establishes compact correlations -
Chemistry determines their shape