Optical UV visible Remote Sensing I Lecture 8 Summer

Optical (UV/visible) Remote Sensing I Lecture 8 Summer 2006 B. -M. Sinnhuber

Material for this lecture by courtesy of Christian von Savigny and Andreas Richter

The optical (UV-visible-NIR) spectral range 1 nm 10 nm 100 nm X-rays EUV 200 nm Vacuum UV 100 nm 400 nm Near UV 700 nm Visible 280 nm 320 nm UV C UV B 5 m NIR 400 nm UV A IR

The sun as a light source • • the solar spectrum can be approximated by a black body at temperature 5780 K absorption in the solar atmosphere leads to Fraunhofer lines in the atmosphere, the solar radiation is attenuated by scattering and absorption strong absorption by O 3, O 2, H 2 O und CO 2 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

$Prism spectrometer The prisms exploits refraction in media with different refractive indices n for$

Prism spectrometer The prisms exploits refraction in media with different refractive indices n for spectral dispersion: Refraction is described by Snell’s law: n = c 0 / c is the refractive index c 0 is the speed of light in vaccum ’ nprism > nmedium

$Diffraction by a grating Gratings are the most common dispersing elements used in remote$

Diffraction by a grating Gratings are the most common dispersing elements used in remote sensing instruments: g g distance between grating grooves g m m diffraction order wavelength For constructive interference, the path difference between two neighboring grating rules has to be a multiple of the wavelength:

Fourier Transform Spectrometers (FTS of FTIR) FTS = Fourier Transform Spectrometer / FTIR = Fourier Transform Infra-Red Spectrometer Michelson interferometer Fixed mirror x L 2/2 Source Beam splitter Movable mirror L 1/2 I(x) Detector 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)

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

Cross secton Principle of wavelength pairs (online - off-line) 1 Complication: Light path dependent on 2 Note: Absorption cross section typically also dependent on x Known from measurements of solar spectrum Absorber column amount along effective light path

Dobson’s spectrophotometer I Quartz plates Adjustable wedge Fixed slits Prisms Detector: photomultiplier Org. photographic plate

Dobson’s spectrophotometer II Used wavelength pairs 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

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 stratosphere - most of the backscattered UV radiation comes from the troposphere - little absorption by ozone occurs in the troposphere - little scattering occurs in the stratosphere - radiation reaching the satellite passes through the ozone layer twice • Backscatter UV measurements allow retrieval of total O 3 columns and also vertical profiles, but with poor vertical resolution (7 – 10 km) • Measure O 3 slant column and use Radiative transfer model to convert to vertical column

Backscatter UV Ozone measurements II 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

Solar occultation measurements I I 0( ) spectrum at the highest tangent altitude with negligible atmospheric extinction I( , zi) spectrum at tangent altitude zi within the atmosphere Lo. S: line of sight With αext, being the total extinction coefficient at position s along the line of sight Lo. S. Extinction is usually due to Rayleigh-scattering, aerosol scattering and absorption by minor constituents:

Solar occultation measurements II Due to the different spectral characteristics of the different absorbers and scatterers the optical depth due to, e. g. , O 3 can be extracted. Absorption cross-section O 3 Number density If we assume that the cross-section does not depend on x, i. e. , not on temperature and/or pressure, then 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 O 3 profile 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 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-3 M 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, H 2 O, NO 2, aerosols GOMOS (Global Ozone Monitoring by Occultation of Stars) on Envisat will performs UV-visible occultation using stars SCIAMACHY (Scanning Imaging Absorption spectro. Meter for Atmospheric CHartograph. Y) on Envisat performs solar and lunar occultation measurements providing e. g. , O 3, NO 2, and (nighttime) NO 3 profiles.

Limb scatter measurements I Modelled limb-radiance profiles At 280 nm Optically thin regime “Knee” Optically thick regime • 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 om limb-scattered f spectra emplying a radiative transfer model (RTM) to simulate the measurements Retrieval without forward model not possible

Limb scatter measurements II Examples: SME (Solar Mesosphere Explorer) launched in 1981, carried the first ever limb scatter satellite instruments. Mesospheric O 3 profiles were retrieved using the Ultraviolet Spectrometer and stratospheric NO 2 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 O 3, NO 2, OCl. O, Br. O SCIAMACHY (Scanning Imaging Absorption Spectro. Meter for Atmospheric CHartograph. Y), launched on Envisat in 2002. Retrieval of vertical profiles of O 3, NO 2, OCl. O, Br. O and aerosols

Advantages and disadvantages of measurement techniques: TECHNIQUE ADVANTAGES DISADVANTAGES Emission • doesn’t require sunlight • long time series • simple retrieval technique • provides global maps twice a day (good spatial coverage) • slightly less accurate than backscatter UV • long horizontal path for limb obs. Backscatter UV • accurate • long time series • good horizontal resolution due to nadir viewing • requires sunlight, so can’t be used at night or over winter poles • poor vertical resolution below the ozone peak (~30 km) due to the effects of multiple scattering and reduced sensitivity to the profile shape Occultation • simple equipment • simple retrieval technique • good vertical resolution • self-calibrating • can only be made at satellite sunrise and sunset, which limits number and location of meas. • long horizontal path Limb Scattering • excellent spatial coverage • good vertical resolution • data can be taken nearly continuously • complex viewing geometry • poor horizontal resolution

Example: Onion peel inversion of occultation observations a 32/2 a 22 a 11 a 21/2 (TH 1) (TH 2) (TH 3) (TH 4) (TH 5) Sun • • • Earth x 5 x 4 x 3 x 2 x 1 xi : O 3 concentration at altitude zi yj : O 3 column density at tangent height THj The matrix elements aij correspond to geometrical path lengths through the layers

Unconstrained least squares solution Inverse of A exists if the determinant of A is not zero Standard approach: least squares solution (assume N=M) Minimize: by: This leads to:

Constrained least squares solution Add additional condition to constrain the solution, e. g. : is a smoothing or constraint coefficient

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 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 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. SZA Offset for clarity only!

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

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: scattering efficiency unattenuated intensity integral over light path absorption by all trace gases j exponential from Lambert Beer’s law extinction by Mie scattering extinction by Rayleigh scattering

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:

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: differential cross-section slant column polynomial

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. intensity with absorption (the measurement result) intensity without or with less absorption (reference measurement) absorption cross-sections (measured in the lab) slant columns SCj are fitted polynomial (bp* are fitted)

Example of DOAS data analysis measurement O 3 optical depth differential optical depth NO 2 residual H 2 O Ring

Example for satellite DOAS measurements • Nitrogen dioxide (NO 2) 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 NO 2 providing the first global maps of NOx pollution • after 10 years of measurements, trends can also be observed A. Richter et al. , Increase in tropospheric nitrogen dioxide over China observed from space, Nature, 437 2005 GOME annual changes in tropospheric NO 2 1996 - 2002