powerpoint presentation · ppt file · web view2008-09-15 · california institute of technology...
TRANSCRIPT
2005 March 15 Oxford Spring School 1
The EOS AURA Tropospheric Emission Spectrometer (TES)1: Principles & Techniques
Reinhard Beer
Science DivisionThe Jet Propulsion Laboratory
California Institute of TechnologyPasadena, CA 91109, USA
2005 March 15 Oxford Spring School 1
Recommended Texts
“Remote Sensing by Fourier Transform Spectrometry”, Reinhard Beer, John Wiley & Sons Inc., New York (1992)
“The Fast Fourier Transform”, E. Oran Brigham, Prentice-Hall Inc., New Jersey (1974)
Both texts are somewhat dated, but still serve as good introductions.
2005 March 15 Oxford Spring School 1
Global measurements of tropospheric ozone and its precursors from TES combined with in-situ data and model predictions will address the following key questions:
How is the increasing ozone abundance in the troposphere affecting
The TES Experiment
- air quality on a global scale?
- climate change?- oxidizing reactions that “cleanse” the atmosphere?
2005 March 15 Oxford Spring School 1
Product Name Required Sensitivity*
Level 1A Interferograms N LLevel 1B Spectral Radiances N LAtmospheric Temperature Profile N L 0.5KSurface Skin Temperature N 0.5KLand Surface Emissivity** N 0.01Ozone (O3 )VMR Profile N L 1 - 20 ppbvWater Vapor (H2O) VMR Profile N L 1 - 200 ppmvCarbon Monoxide (CO) VMR Profile N L 3 - 6 ppbvMethane (CH4) VMR Profile N L 14 ppbvNitrogen Dioxide (NO2) VMR Profile L 15 - 25 pptvNitric Acid (HNO3) VMR Profile L 1 - 10 pptvNitrous Oxide (N2O) VMR Profile N L Control***
*** Tropospheric concentration known
Product Source
* Sensitivity range maps to expected concentration range** Water emissivities are known and are therefore input , not output, parameters
TES Standard ProductsN = Nadir, L = Limb Viewing
2005 March 15 Oxford Spring School 1
The TES Instrument
TES is an infrared imaging Fourier Transform Spectrometer (FTS) operating in the spectral range 650 – 3050 cm-1 (roughly 3.3 – 15.4 m).
It features 4 1x16 element optically-conjugated focal plane arrays each optimized for a different spectral region and operating at a temperature of 63K using mechanical refrigerators. In addition, each focal plane is equipped with interchangeable cooled filters that limit the instantaneous spectral bandwidths to about 250 cm-1. This provides needed control over the instrument thermal background and reduces the data rate.
Except for two external mirrors (part of the pointing system), the entire optical path is radiatively-cooled to about 165K, further to reduce the instrument background.
A particular feature of TES is that it can, unlike any other spaceborne FTS instrument, observe both downwards and sideways to the limb.
2005 March 15 Oxford Spring School 1
Earth
Atmosphere
2005 March 15 Oxford Spring School 1
First Principles
2005 March 15 Oxford Spring School 1
The Principle of Superpositionelectromagnetic amplitudes are additive
Imagine 2 wavetrains of the same amplitude and frequency:
Phase in Wavelengths
0 1 2 3 4
Am
plitu
de
+A
0
-A
2005 March 15 Oxford Spring School 1
Let the wavetrains be a A x1 2 co s( ) and a A x2 22 cos( )
Upon addition, we get a a a A x x A x x x x 1 2 1 2 1 2 1 22 2 2[cos( cos( ] co s( { } ) cos( { })
Setting x x1 2 and dropping the subscripts we end with
a A x 2 2 2cos( cos( { / } )
which represents a wavetrain of the same frequency but phase shifted by /2 (second cosine term) and an amplitude (first cosine term) that depends on the phase difference. Observe that the resultant amplitude can vary between 2A and zero.
2005 March 15 Oxford Spring School 1
Spectrometric Units
The units most commonly used in spectrometry are the nanometre (1 nm = 10-9 metres), the micrometre (1 m = 10-6 metres) and the reciprocal centimetre (cm-1).
For reasons that will become clear, it is this latter unit that is most commonly used in infrared spectrometry.
CONVERSION FACTORS:1 cm-1 = (Frequency in Hertz)/(Velocity of light) = 107/(wavelength in nanometres)
= 104/(wavelength in micrometres)
Occasionally, one will see the cm-1 referred to as the wavenumber.
This is a description, NOT A UNIT!
2005 March 15 Oxford Spring School 1
Two-Beam Interference
Recall the expression a A x 2 2 2co s( cos( { / } )
The intensity is therefore I a A 2 2 24 cos (
Replacing A2 by Iin, the incident intensity of each wavetrain, we get
I I Iin in 4 1 22cos ( [ cos( ]
It is evident that I will alternate between 4Iin and 0 as changes through an amount 1 /
These two extreme conditions are termed constructive anddestructive interference, respectively (although any intermediate condition is equally possible).
2005 March 15 Oxford Spring School 1
Source
Input Frequency
Detector
M oving ArmVelocity V cm /sec
Condenser
Collim ator
PRINCIPLE OF THE MICHELSON INTERFEROMETER
cm -1Beamsplitt
er
Output Frequency = 2.V. Hz
2005 March 15 Oxford Spring School 1
K
I
J
d
Geometric Equivalent of a Michelson Interferometer
M irror 1 Im age of M irror 2Beam 2
Beam 1 The geometric path differencebetween beams 1 & 2 is IJ + JK
where
IJ+JK = d/cos( +d.cos(2./cos( = 2.d.cos()
This expression can also be described by a number of wavelengths: n = 2.d.cos().If the interferometer is embedded in a medium of refractive index , this becomes n = 2..d.cos().
As stated earlier, we prefer to work in a frequency unit = 1/ so, finally, we get
n = 2...d.cos()
which is the fundamental equation for most types of interferometer.
2005 March 15 Oxford Spring School 1
Spectral Resolution (1)
Spectral Resolution is a measure of the ability of the spectrometer to discriminate between adjacent spectral features.
As we shall see, the spectral resolution of a Fourier Transform Spectrometer (henceforth FTS) is inversely proportional to the maximum path difference employed.
It must, however, be recognized that improved spectral resolution comes at the price of reduced signal-to-noise ratio, so the maximum path difference employed must be carefully “tuned” to the specific problem to be addressed. In short – enough but not too much!
2005 March 15 Oxford Spring School 1
Normalized Shapes of Doppler & Lorentz Lines
Normalized Frequency, ( )/
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Nor
mal
ized
Abs
orpt
ion
Coe
ffici
ent
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Doppler
Lorentz
2005 March 15 Oxford Spring School 1
The 1976 US Standard Atmosphere Temperature Profile.Zones important to TES are indicated
Temperature, kelvins
210 215 220 225 230 235 240 245 250 255 260 265 270 275 280 285 290
Alti
tude
, km
0
5
10
15
20
25
30
35
40
45
50
Troposphere
Tropopause
Stratosphere
2005 March 15 Oxford Spring School 1
Weak Line Full-Width-at-Half-Maximum, cm-1
0.0001 0.001 0.01 0.1 1
Alti
tude
, km
0
5
10
15
20
25
30
35
40
45
50
Range for typicalLorentz lines
Range for typical Dopplerlines @ 1000 cm-1
TES
Lim
b Vi
ewin
g
TES
Nad
ir Vi
ewin
g
Variation of Typical Linewidths with Altitude
2005 March 15 Oxford Spring School 1
Spectral Resolution (2)Now understanding what is desirable for spectral resolution, how does this apply to an FTS? Remembering an earlier equation for the monochromatic case:
I I Iin in 4 1 22cos ( [ cos( ]
we can readily change it (and some notation) for the broad-band case:
I x F d F x d( ) ( ) ( ) cos( )
0 02
The first integral is simply the area under the spectrum (i.e., is an additive constant that can be ignored). The second integral will be recognized as the cosine Fourier Transform of F ( )
Now, Fourier Transforms are defined over infinite ranges in both domains. We overcome the problem in the frequency domain through the use of optical filters. That is, we forceoutside some range In the path difference (x) domain, it is a different story.
F ( ) 0
Because we can only change path difference (scan) over some finite range Xmax (in either sense, or both), this is mathematically equivalent to multiplying the “infinite” interferogram by a boxcar function that = 1 over the scanned range and is zero elsewhere.
Now, a multiplication in interferogram space is identical to a convolution in spectrum space. The Fourier Transform of a boxcar is the well-known “sinc” (= sin[x]/x) function which therefore serves to define the spectral resolution of an ideal FTS.
2005 March 15 Oxford Spring School 1
The sinc (= sin(x)/x) function
Frequency offset in cm-1 * Maximum path difference in cm
-5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Am
plitu
de
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
2005 March 15 Oxford Spring School 1
Reality
2005 March 15 Oxford Spring School 1
2005 March 15 Oxford Spring School 1
View of the TES engineering model interferometer retroreflector
2005 March 15 Oxford Spring School 1
2A3 1A1 1A2
2005 March 15 Oxford Spring School 1
2005 March 15 Oxford Spring School 1
TES Instrumental Line Shape (ILS)CO2 Laser 0001 - 0200 P(20) line (nominally @ 1046.8543 cm-1) Model prediction
Frequency, cm-1
1046.6 1046.7 1046.8 1046.9 1047.0 1047.1 1047.2
Rel
ativ
e am
plitu
de
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0181 cm-1
2005 March 15 Oxford Spring School 1
+ Z
+ X
+ YA u ra S p a c ec ra ft C o ord in a te S y stem
T E SL im b V ie w -64 .2 o
S p ac e V ie w -7 2 o
~ 3 5 k mG eo id S u rfa c e
N a d ir
+ 4 5 In -tra cko
-45 In -tracko
-45 C ross -tr ack
o
+ 4 5 C ro ss -tr ack
o
T E S
F ligh t P a th ~3 50 k m
16 k m a im p oin t
TES Observation Geometry
2005 March 15 Oxford Spring School 1
TES Operating ModesGlobal Survey: 16 orbits of nadir & limb observations repeated every other day. This is the source of Standard Products.
Stare: Point at a specific latitude & longitude for up to ~4 minutes.
Transect: Point at a set of contiguous latitudes & longitudes to cover ~850 km.
Step-&-Stare: Point at nadir for 4 seconds (5.2 seconds with necessary reset). Spacecraft has moved ~40 km. Point at nadir again. Repeat indefinitely.
Limb Drag: Point at the trailing limb (16 second scans). Repeat indefinitely.
These last 4 modes constitute Special Products that are obtained only when no Global Survey is scheduled.
2005 March 15 Oxford Spring School 1
2005 March 15 Oxford Spring School 1
TES Step & Stare, Stare and Transect Modes
x x
STEP & STARE: A set of nadir footprints spaced about 35 km apart. Can be indefinite.
STARE: Point at a specific latitude & longitude for up to 210 seconds.
TRANSECT: A set of exactly contiguous latitudes & longitudes in a line up to 885 km long.
2005 March 15 Oxford Spring School 1
•Fourier transform spectrometer Fourier transform spectrometer •Wavelength response: 3.3 to 15.4 micronWavelength response: 3.3 to 15.4 micron•One scan every 4 or 16 sec. (0.1 cmOne scan every 4 or 16 sec. (0.1 cm-1-1 or 0.025 cm or 0.025 cm-1-1 res.) res.)•Four optically-conjugated 1x16 pixel detector arraysFour optically-conjugated 1x16 pixel detector arrays•Spatial resolution of 5 x 8 km at nadir & 2.3 km at limbSpatial resolution of 5 x 8 km at nadir & 2.3 km at limb•Passively cooledPassively cooled•2-axis gimbaled pointing mirror2-axis gimbaled pointing mirror
TES specifications
2005 March 15 Oxford Spring School 1
TES
Nadir View
Limb View
TES on the Aura Spacecraft
2005 March 15 Oxford Spring School 1
AURA Launch; 2004 July 15, 03:02 PDT
2005 March 15 Oxford Spring School 1
Level 1A: Produces geolocated interferograms.
TES Algorithm Overview
Level 1B: Produces radiometrically and frequency calibrated spectra with NESR.
Level 2: Produces VMR and temperature profiles.
Level 3: Produces global maps.
2005 March 15 Oxford Spring School 1
CONCLUSIONS
TES is fulfilling its promise to provide the first-ever global overview of the key constituents of tropospheric chemistry
and their inter-regional transport
For more information:
http://tes.jpl.nasa.gov