Optical (UV/visible) Remote Sensing

Optical (UV/visible) Remote Sensing

Remote Sensing ILecture 8

Summer 2006

B.-M. Sinnhuber

Material for this lecture by courtesy ofChristian von Savigny and Andreas Richter

The optical (UV-visible-NIR) spectral range

1 nm 10 nm 100 nm 200 nm 400 nm 700 nm

5 m

VisibleVacuum UV Near UV NIR IREUVX-rays

100 nm 400 nm320 nm280 nm

UV AUV C UV B

The sun as a light source

• the solar spectrum can be approximated by a black body at temperature 5780K

• absorption in the solar atmosphere leads to Fraunhofer lines

• in the atmosphere, the solar radiation is attenuated by scattering and absorption

• strong absorption by O3, O2, H2O und CO2

• there are some atmospheric windows where absorption is small

• multitude of Fraunhofer lines• 11 year solar cycle, particularly relevant at

short wavelengths < 300 nm• spectrum varies over the solar disk• Doppler shift resulting from rotation of sun• variation of intensity due to changes in

distance sun - earth

=> sun is not an ideal light source!

Instrumentation for remote sensing measurements

Atmospheric remote sensing methods usually require spectrally resolved radiation measurements

spectrally dispersing elements required

The standard radiation-dispersing devices are:

•Prisms•Gratings•Michelson Interferometers•Fabry-Perot Interferometers

Prism spectrometer

The prisms exploits refraction in media with different refractive indices n for spectral dispersion:

’

nprism > nmedium

Refraction is described by Snell’s law:

n

n

sin

sin

n = c0 / c is the refractive indexc0 is the speed of light in vaccum

For constructive interference, the path difference between two neighboring grating rules has to be a multiple of the wavelength:

Diffraction by a grating

Gratings are the most common dispersing elements used in remote sensing instruments:

g

m

g

g distance between grating grooves

m diffraction order

wavelength

g

mλsinαm

Fourier Transform Spectrometers (FTS of FTIR)

Michelson interferometer

Measured is the intensity of the two interfering light beams as a function of the position x of the movable mirror: I(x) is called interferogram

The spectrum S() is the Fourier-transform of I(x)

x

Movable mirror

Fixed mirror

Source

Beam splitter

Detector

L1/2

L2/2

I(x)

FTS = Fourier Transform Spectrometer / FTIR = Fourier Transform Infra-Red Spectrometer

Monochromators and spectrometers

Monochromators are single color, tunable optical band pass filters

Spectrometers measure a continuous spectral range simultaneously

Note: Depending on the type of detector, a prism or grating instrument can be a monochromator or a spectrometer

Principle of wavelength pairs (online - off-line)

C

ross

sec

ton

1 2

dxxndxxn

eI

I

I

I 21

20

10

2

1

dxxn

eII2

202

dxxn

eII1

101

dxxn

eI

I

I

I 21

20

10

2

1

dxxn

I

I

I

I21

20

10

2

1 lnln

Note: Absorption cross section typically also dependent on x

Complication: Light path dependent on

Known from measurements of solar spectrum

Absorber column amount along effective light path

Dobson’s spectrophotometer I

Quartz platesAdjustable wedgeFixed slitsPrismsDetector: photomultiplier

Org. photographic plate

Dobson’s spectrophotometer II

Name WL 1 [nm] WL 2 [nm]

A 305.5 325.4

B 308.8 329.1

C 311.45 332.4

D 317.6 339.8

C’ 332.4 453.6

Longest existing ozone time series

Used wavelength pairs

Overview of satellite observations geometries

Measured signal:Reflected and scattered sunlight

Measured signal:Directly transmitted solar radiation

Measured signal:Scattered solar radiation

Backscatter UV Ozone measurements I

• Solar UV radiation reflected from the surface and backscattered by atmosphere or clouds is absorbed by ozone in the Hartley-Huggins bands (< 350 nm)

• Note:most ozone lies in the stratospheremost of the backscattered UV radiation comes from the tropospherelittle absorption by ozone occurs in the tropospherelittle scattering occurs in the stratosphereradiation reaching the satellite passes through the ozone layer twice

• Backscatter UV measurements allow retrieval of total O3 columns and also vertical profiles, but with poor vertical resolution (7 – 10 km)

• Measure O3 slant column and use Radiative transfer model to convert to vertical column

Examples:

BUV (Backscatter Ultraviolet) instrument on Nimbus 4, 1970-1977

SBUV (Solar Backscatter Ultraviolet) instrument on Nimbus 7, operated from 1978 to 1990

SBUV/2 (Solar Backscatter Ultraviolet 2) instrument on the NOAA polar orbiter satellites: NOAA-11 (1989 -1994), NOAA-14 (in orbit) can measure ozone profiles as well as columns

TOMS (Total Ozone Mapping Spectrometer) first on Nimbus 7, operated from 1978 to 1993. Then three subsequent versions: Meteor 3 (1991-1994), ADEOS (1997), Earth Probe (1996-). Measures total ozone columns.

GOME (Global Ozone Monitoring Experiment) launched on ESA's ERS-2 satellite in 1995 employs a nadir-viewing BUV technique that measures radiances from 240 to 793 nm. Measures O3 columns and profiles, as well as columns of NO2, H2O, SO2, BrO, OClO.

Backscatter UV Ozone measurements II

Solar occultation measurements I

I0() spectrum at the highest tangent altitude with negligible atmospheric extinction

I(,zi) spectrum at tangent altitude zi within the atmosphere

λI

THλ,I)THτ(λ,-exp

0

ii

)LoS(TH

λext,

i

dxxexp LoS: line of sight

With αext, being the total extinction coefficient at position s along the line of sight LoS.

Extinction is usually due to Rayleigh-scattering, aerosol scattering and absorption by minor constituents:

xxxx gasesλ

aerosolλ

Rayleighλλext,

Disadvantage of solar occultation measurements:

The occultation condition has to be met: Measurements only possible during orbital sunrises/sunsets

For typical Low Earth Orbits there are 14–15 orbital sunrises and sunsets per day

poor geographical coverage

Solar occultation measurements II

If we assume that the cross-section does not depend on x, i.e., not on temperature and/or pressure, then

iO

THLoS

Oiozone THcσdxxnσTHλ,τ

3

i

3 With the column density c(zi)

The measurement provides a set of column densities integrated along the line of sight for different tangent altitudes zi.

Inversion to get vertical O3 profile

Due to the different spectral characteristics of the different absorbers and scatterers the optical depth due to, e.g., O3 can be extracted.

LoS

Oiozone dxxnxσTHλ,τ

3

Absorption cross-section

O3 Number density

Solar occultation measurements III

Examples:

SAGE (Stratospheric Aerosol and Gas Experiment) Series provided constinuous observations since 1984 to date

Latest instrument is SAGE III on a Russian Meteor-3M spacecraft

HALOE (Halogen Occultation Experiment) on UARS (Upper Atmosphere Research Satellite) operated from 1991 until end of 2005, employing IR wavelengths

POAM (Polar Ozone and Aerosol Measurement) series use UV-visible solar occultation to measure profiles of ozone, H2O, NO2, aerosols

GOMOS (Global Ozone Monitoring by Occultation of Stars) on Envisat will performs UV-visible occultation using stars

SCIAMACHY (Scanning Imaging Absorption spectroMeter for Atmospheric CHartographY) on Envisat performs solar and lunar occultation

measurements providing e.g., O3, NO2, and (nighttime) NO3 profiles.

Limb scatter measurements I

• Radiation is detected which is scattered into the field of view of the instrument along the line of sight and also transmitted from the scattering point to the instrument

• Solar radiation pickts up absorption signatures along the way

• Also: Light path can be modified by atmospheric absorption

• Vertical profiles of several trace constituents can be retrieved fom limb-scattered spectra emplying a radiative transfer model (RTM) to simulate the measurements

Retrieval without forward model not possible

0

10

20

30

40

50

60

70

80

90

10-6

10-5

10-4

10-3

10-2

10-1

305 nm

280 nm

Limb radiance [a.u.]

Tang

ent h

eigh

t [km

] “Knee”

Optically thin regime

Optically thick regime

At 280 nmModelled limb-radiance profiles

Examples:

SME (Solar Mesosphere Explorer) launched in 1981, carried the first ever limb scatter satellite instruments. Mesospheric O3 profiles were retrieved using the Ultraviolet Spectrometer and stratospheric NO2 profiles were retrieved using the Visible Spectrometer

MSX satellite – launched in 1996 , carried a suite of UV/visible sensors (UVISI)

SOLSE (Shuttle Ozone Limb Sounding Experiment) flown on the Space Shuttle flight in 1997. Provided good ozone profiles with high vertical resolution down to the tropopause

OSIRIS (Optical Spectrograph and Infrared Imager System) launched in 2001 on Odin satellite. Retrieval of vertical profiles of O3, NO2, OClO, BrO

SCIAMACHY (Scanning Imaging Absorption SpectroMeter for Atmospheric CHartographY), launched on Envisat in 2002. Retrieval of vertical profiles of O3, NO2, OClO, BrO and aerosols

Limb scatter measurements II

TECHNIQUE ADVANTAGES DISADVANTAGES

Emission • doesn’t require sunlight • slightly less accurate than• long time series backscatter UV• simple retrieval technique • long horizontal path for limb obs.• provides global maps twicea day (good spatial coverage)

Backscatter UV • accurate • requires sunlight, so can’t be• long time series used at night or over winter poles• good horizontal resolution • poor vertical resolution below thedue to nadir viewing ozone peak (~30 km) due to the

effects of multiple scattering andreduced sensitivity to the profile

shape

Occultation • simple equipment • can only be made at satellite• simple retrieval technique sunrise and sunset, which limits• good vertical resolution number and location of meas.• self-calibrating • long horizontal path

Limb Scattering • excellent spatial coverage • complex viewing geometry• good vertical resolution• data can be taken nearly • poor horizontal resolutioncontinuously

Advantages and disadvantages of measurement techniques:

Example: Onion peel inversion of occultation observations

Earth x1x2x3x4x5

(TH1)(TH2)(TH3)(TH4)(TH5)

•

•

•

My

y

y

y

y

3

1

1

Nx

x

x

x

x

3

1

1

xy

A

xi : O3 concentration at altitude zi

yj : O3 column density at tangent height THj

MNMM

N

N

aaa

aaa

aaa

21

22221

11211

A

a11 a21/2a22

a32/2

The matrix elements aij correspond to geometrical path lengths through the layers

N

jjiji xay

1

Sun

yx -1A

Inverse of A exists if the determinant of A is not zero

Standard approach: least squares solution (assume N=M)

Mixay j

N

jijii ,,1,

1

Minimize:

by:

M

i

M

i

N

jijiji yxa

1 1

2

1

22

01

2

1

M

i

N

jijij

k

yxax

This leads to:

01 1

M

iik

N

jijij ayxa

yx TT AAA

yx TT AAA

-1

Unconstrained least squares solution

Constrained least squares solution

M

i

N

jjj

M

i

N

jijij

N

jji xxyxaxx

1

2

11

1

2

1

2

1

2

Add additional condition to constrain the solution, e.g.:

is a smoothing or constraint coefficient coefficient

02

11

1

2

1

N

jjj

M

i

N

jijij

k

xxyxax

02 111 1

kkk

M

iik

N

jijij xxxayxa

0 xyx

HAAA TT yx TT AHAA

-1

Differential Optical Absorption Spectroscopy (DOAS)

• remote sensing measurement of atmospheric trace gases in the atmosphere• measurement is based on absorption spectroscopy in the UV and visible wavelength

range• to avoid problems with extinction by scattering or changes in the instrument throughput,

only signals that vary rapidly with wavelength are analysed (thus the differential in DOAS)

• measurements are taken at moderate spectral resolution to identify and separate different species

• when using the sun or the moon as light source, very long light paths can be realised in the atmosphere which leads to very high sensitivity

• even longer light paths are obtained at twilight when using scattered light

• scattered light observations can be taken at all weather conditions without significant loss in accuracy for stratospheric measurements

• use of simple, automated instruments for continuous operation

Multiple light paths

SZA

Offset for clarity only!

In practice, many light paths through the atmosphere contribute to the measured signal.

Intensity measured at the surface consist of light scattered in the atmosphere from different altitudes

For each altitude, we have to consider• extinction on the way from the top

of the atmosphere• scattering probability• extinction on the way to the surface

in first approximation, the observed absorption is then the absorption along the individual light paths weighted with the respective intensity.

Airmass factors

VCSC

AMF,...),(

The airmass factor (AMF) is the ratio of the measured slant column (SC) to the vertical column (VC) in the atmosphere:

The AMF depends on a variety of parameters such as• wavelength• geometry• vertical distribution of the species• clouds• aerosol loading• surface albedo

The basic idea is that the sensitivity of the measurement depends on many parameters but if they are known, signal and column are proportional

VCSC

Airmass factors: dependence on solar zenith angle (SZA)For a stratospheric absorber, the AMF

strongly increases with solar zenith angle (SZA) for ground-based, airborne and satellite measurements.

Reason: increasing light path in the upper atmosphere (geometry)

For an absorber close to the surface, the AMF is small, depends weakly on SZA but at large SZA rapidly decreases.

Reason: light path in the lowest atmosphere is short as it is after the scattering point for zenith observation.

=> stratospheric sensitivity is highest at large SZA (twilight)=> tropospheric sensitivity is largest at high sun (noon)=> diurnal variation of slant column carries information on vertical profile

Airmass factors: dependence on absorber altitude

• The AMF depends on the vertical profile of the absorber. The shape of the vertical dependence depends on wavelength, viewing geometry and surface albedo.

• For zenith-viewing measurements, the sensitivity increases with altitude (geometry).

• For satellite nadir observations, the sensitivity is low close to the surface over dark surfaces (photons don’t reach the surface) but large over bright surfaces (multiple scattering).

=> the vertical profile must be known for the calculation of AMF

Airmass factors: dependence on wavelength

• the AMF depends on wavelength as Rayleigh scattering is a strong function of wavelength and also the absorption varies with wavelength

• at low sun, the AMF is smaller in the UV than in the visible as more light is scattered before travelling the long distance in the atmosphere.

• at high sun, the opposite is true as a result of multiple scattering

• UV measurements are more adequate for large absorption

• in the case of large absorptions, the nice separation of fit and radiative transfer is not valid anymore as AMF and absorption are correlated

• different wavelengths “see” different parts of the atmosphere which can be used for profile retrieval

DOAS equation I

The intensity measured at the instrument is the extraterrestrial intensity weakened by absorption, Rayleigh scattering and Mie scattering along the light path:

}))()()()()()((exp{)(),(),(1

0

dssssIaI RayRay

J

jMieMiejj

absorption by all trace gases j

extinction by Mie scattering

extinction by Rayleigh scattering

unattenuated intensity

integral over light pathscattering efficiency

exponential from Lambert Beer’s law

DOAS equation II

if the absorption cross-sections do not vary along the light path, we can simplify the equation by introducing the slant column SC, which is the total amount of the absorber per unit area integrated along the light path through the atmosphere:

}))()()()()()((exp{)(),(),(1

0

dssssIaI RayRay

J

jMieMiejj

dssSC jj )(

})()()(exp{)(),(),(1

0 RayRay

J

jMieMiejj SCSCSCIaI

DOAS equation III

As Rayleigh and Mie scattering efficiency vary smoothly with wavelength, they can be approximated by low order polynomials. Also, the absorption cross-sections can be separated into a high (“differential”) and a low frequency part, the later of which can also be included in the polynomial:

})()()(exp{)(),(),(1

0 RayRay

J

jMieMiejj SCSCSCIaI

})('exp{)(),(),(1

0

J

j p

ppjj bSCIaI

4 Ray

Mie

20 ' low

polynomial

differential cross-section

slant column

DOAS equation IV

Finally, the logarithm is taken and the scattering efficiency included in the polynomial. The result is a linear equation between the optical depth, a polynomial and the slant columns of the absorbers. by solving it at many wavelengths (least squares approximation), the slant columns of several absorbers can be determined simultaneously.

J

j p

ppjj bSCII

1

*0 )('))(/),(ln(

polynomial (bp* are fitted) slant columnsSCj are fitted

absorption cross-sections (measured in the lab)

intensity with absorption (the measurement result)

intensity without or with less absorption (reference measurement)

Example of DOAS data analysis

measurement optical depth differential optical depth

O3

H2O

NO2

residual

Ring

Example for satellite DOAS measurements

• Nitrogen dioxide (NO2) and NO are key species in tropospheric ozone formation

• they also contribute to acid rain• sources are mainly

anthropogenic (combustion of fossil fuels) but biomass burning, soil emissions and lightning also contribute

• GOME and SCIAMACHY are satellite borne DOAS instruments observing the atmosphere in nadir

• data can be analysed for tropospheric NO2 providing the first global maps of NOx pollution

• after 10 years of measurements, trends can also be observed

1996 - 2002

GOME annual changes in tropospheric NO2

A. Richter et al., Increase in tropospheric nitrogen dioxide over China observed from space, Nature, 437 2005