1 Satellite Remote Sensing I - PowerPoint Presentation

Slide1

1

Satellite Remote

Sensing I

UV/vis observations of the troposphere

Lecture at the

PANDA summer school 2015

Bremen, August 25, 2015

Andreas Richter

Institute of Environmental Physics

University of Bremen

Bremen, Germany

(

richter@iup.physik.uni-bremen.de

)

Slide2

2

Overview

What is Remote Sensing?

How can trace gases be detected using UV/vis remote sensing from satellite ?

What are typical light paths for UV/vis satellite observations?

What are the sensitivities of the measurements?

What needs to be considered when comparing satellite data to other observations / to models?

Some examples of UV/vis instruments & applications

Slide3

Who am I?Leading DOAS group at Institute of Environmental Physics, University of Bremen, head: Prof. John Burrows.Working on all aspects of UV/vis remote sensingInstruments (ground, airborne, satellite)Radiative transferRetrieval algorithmsData interpretationSome atmospheric topics I‘m interested inNOx emissions, distributionsand chemistryEmission changes / trendsHalogen chemistry inpolar regions3

www.doas-bremen.de

Slide4

Basic Principles of Remote Sensing“Remote sensing is the science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with the object, area, or phenomenon under investigation“ (Lillesand and Kiefer 1987)“The art of dividing up the world into little multi-coloured squares and then playing computer games with them to release unbelievable potential that's always just out of reach.” (Jon Huntington, Commonwealth Scientific and Industrial Research Organisation Exploration, Geoscience, Australia)

4

Slide5

Schematic of Remote Sensing Observation

5

Validation

Sensor

Measurement

Object

Changed Radiation

Radiation

Data Analysis

Final Result

A priori information

Forward Model

Slide6

The Electromagnetic Spectrumnearly all energy on Earth is supplied by the sun through radiationwavelengths from many meters (radio waves) to nm (X-ray) short wavelength = high energyradiation interacts with atmosphere and surfaceabsorption (heating, shielding)excitation (energy input, chemical reactions)re-emission (energy balance)

6

Slide7

Wavelength Ranges in Remote SensingUV: some absorptions + profile information aerosols vis: surface information (vegetation) some absorptions aerosol information IR: temperature information cloud information water / ice distinction many absorptions / emissions + profile information

MW:

no problems with clouds

ice / water contrast

surfaces

some emissions + profile information

7

Slide8

Wavelength Ranges in Remote SensingUV: some absorptions + profile information aerosols vis: surface information (vegetation) some absorptions aerosol information IR: temperature information cloud information water / ice distinction

many absorptions / emissions

+ profile information

MW:

no problems with clouds

ice / water contrast

surfaces

some emissions + profile information

8

Slide9

How are trace gases detected in the UV/vis?Method used is Absorption SpectroscopyRadiation coming from the sun is attenuated as it passes through the atmosphereThe wavelength dependence of the absorption is molecule specificThe magnitude of the absorption depends on the amount of absorber along the light pathThis is described by Lambert Beer‘s law:ρ = density of absorber λ = wavelength σ = absorption cross-section of absorber9

Slide10

Atmosphere

Absorption on the ground

Scattering / Reflection on the ground

Emission from

the ground

Scattering from a cloud

Transmission through a cloud

Transmission through a cloud

Scattering / reflection on a cloud

Scattering within a cloud

Cloud

Emission from

a cloud

Absorption

Scattering

Aerosol / Molecules

Emission

Radiative

Transfer in the

Atmosphere

passive sensor, UV / visible / IR

10

Slide11

Atmosphere

Absorption on the ground

Scattering / Reflection on the ground

Scattering from a cloud

Transmission through a cloud

Transmission through a cloud

Scattering / reflection on a cloud

Scattering within a cloud

Cloud

Absorption

Scattering

Aerosol / Molecules

Radiative Transfer in the

Atmosphere

passive sensor, UV / visible

11

Slide12

Radiative transfer in the atmosphereThe change of radiation along the light path: Solution is complex, in particular in spherical geometryComplications are polarisation, inealasic

scattering

, angular

dependent surface

reflectivity

,

complex

phase functions of aerosols and cloud particlesEven

if

only

one

scattering

event

is considered,

manypossible light paths need tobe

included 12

Slide13

Retrieval ApproachIn general, a retrieval always consists of Solutions are for example using the Optimal Estimation methodOften, several pieces of information can be extracted such as concentrations in two layers A widely used method which is fast and simple is to separate the retrieval of the amount of absorber found in the measurement andThe length of the light pathThis is often done using the Differential Optical Absorption Spectroscopy (DOAS) 13

Measurement

Forward Model

A priori

Parameters

Simulation

OK?

Slide14

14

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:

absorption by all trace gases j

extinction by Mie scattering

extinction by Rayleigh scattering

unattenuated intensity

integral over light path

scattering efficiency

exponential from Lambert Beer’s law

Slide15

15

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.

Slant Column

=

absorber

concentration

integrated

along

light

path

Slide16

16

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:

polynomial

slant column

differential cross-section

Slide17

17

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.

DOAS retrieves absorber amount along light path – but what is the light path?

Slide18

Satellite Viewing GeometriesDirect Sun / MoonLot‘s of lightSimple geometryVertical resolutionNo tropospheric observationsLow spatial resolutionFew measurementsLimb scattered lightVertical resolutionMany measurementsComplex light pathNo tropospheric observationsLow spatial resolutionNadir scattered light

High spatial resolution

Sensitivity to the troposphere

Little vertical resolution

Complex light path

18

Slide19

Typical light paths: UVDark surfaceStrong Rayleigh scatteringMost photons are scattered above absorption layer=> Low sensitivity to BL signals!

19

sensitivity

altitude

Slide20

Typical light paths: visibleBrighter surfaceSignificant Rayleigh ScatteringMany photons are scattered above absorption layer=> Reduced sensitivity to BL signals!

20

sensitivity

altitude

Slide21

Bright surface (snow, ice): UV and visibleSurface reflection dominatesMultiple scattering in surface layer=> Enhanced sensitivity to BL signals!

21

sensitivity

altitude

Slide22

Typical light paths: NIRBright surface (except for oceans)Negligible Scattering=> Very good sensitivity to BL signals!

22

sensitivity

altitude

Slide23

Typical light paths: visible with cloud

23

sensitivity

altitude

Slide24

Typical light paths: visible with cloudEven if cloud covers only part of the scene, most photons are reflected thereSensitivity below cloud is reducedSensitivity above cloud is enhanced=> Cloud fraction and cloud height are important!

24

sensitivity

altitude

Slide25

Tropospheric sensitivity of UV/vis observationsWavelengthSensitivity decreases towards UVVertical profile of absorberSensitivity decreases to surface unless it is very brightCloudsSensitivity below clouds is close to 0Thin clouds (or aerosols) can increase / decrease sensitivitySurface albedoSensitivity increases with surface reflectanceSolar zenith angleLower sun increases sensitivity geometrically, at least in mid and upper troposphereLower sun decreases sensitivity to the surface layers25

Slide26

Vertical sensitivity of satellite measurementsThe sensitivity of the satellite measurements depends on the altitude of the absorbing layerThis is often expressed in the form of weighting functions which give the sensitivity of the signal as function of altitude of the trace gas layerAs the vertical distribution can usually not be (completely) determined from the measurements, a priori information is needed in the retrievalThe dependence of the retrieved quantity on the real atmospheric profile depends on both, the sensitivity of the measurements and the assumptions made in the a prioriThis is often expressed as averaging kernels which describe the sensitivity of the retrieved quantity on the amounts of trace gas in the different altitudes in the atmosphereComparison of satellite retrievals with other measurements are only meaningful if the averaging kernels are accounted for

26

Slide27

Satellite Orbits(Near) Polar Orbit:orbits cross close to the poleglobal measurements are possiblelow earth orbit LEO (several 100 km)ascending and descending branchspecial case: sun-synchronous orbit:overpass over given latitude always at the same local time, providing similar illuminationfor sun-synchronous orbits: day and night branchesGeostationary Orbit:satellite has fixed position relative to the Earthparallel measurements in a limited area from low to middle latitudes36 000 km flight altitude, equatorial orbit

http://www2.jpl.nasa.gov/basics/bsf5-1.htm

http://www.ccrs.nrcan.gc.ca/ccrs/learn/tutorials/fundam/chapter2/chapter2_2_e.html

27

Slide28

Considerations for use of UV/vis satellite dataSpatial resolutionSatellite data is coarse in comparison to in-situ observations but finer than global models => averaging is neededSpatial samplingDue to orbital pattern, scan and cloud gaps, only part of the globe is covered by observations every day=> other data must be sampled accordinglyTime of overpassSun sun-sychronous orbits have overpasses at a constant local time => other data needs to be interpolated to that timeVertical resolutionSatellite data have coarse or no vertical resolution=> other data must be integrated to become comparableImpact of a priori

The vertical sensitivity of the retrievals (averaging kernel) needs to be considered

Summary: We should always compare apples to apples, not oranges!

28

Slide29

Effect of Spatial Resolution: Example NO2

29

For species with short atmospheric life time, horizontal variability is large

Spatial resolution of sensor is relevant for interpretation

Spatial resolution also influences cloud fraction

Time of overpass may also play a role!

OMI: 13:30 LT

Slide30

Examples30

Slide31

TOMSInstrument: UV discrete (6) wavelengths grating spectrometerOperational: October 1978 - 2004Spatial resolution: 50 x 50 km2Global coverage: 1.5 daysSpecies: O3, SO2

31

Slide32

TOMS: Ozone columns Large scale tropospheric ozone patterns retrieved using the cloud slicing methodDuring El Nino year, clear ozone maximum over IndonesiaOrigins: photochemical smog from biomass burning and change in circulation pattern

32

Ziemke, J. R et al., (2001), “Cloud slicing”: A new technique to derive upper tropospheric ozone from satellite measurements,

J. Geophys. Res.

, 106(D9), 9853–9867

October

1996

October

1997

October

1998

Slide33

GOME / GOME-2Instrument: 4 channelUV/vis grating spectrometerOperational on ERS-2 7.1995 – 6.2003 - 2011Spatial resolution 320 x 40 km2Global coverage: 3 daysSpecies: O3, NO2, HCHO, CHOCHO, BrO, IO, SO2, H2O

33

GOME-2 A

on

MetOp

A since

1.2007

80 x 40

km

2

1.5

days

GOME-2 B

on

MetOp

B since 1.2013

80 x 40 km

2

1.5 days

40

x 40

km

2

3 days

Slide34

34

Satellite NO

2

Trends

NO

2

reductions in Europe and parts of the US

strong increase over China

consistent with significant NO

x

emission changes

7 years of GOME satellite data

DOAS retrieval + CTM-stratospheric correction

seasonal and local AMF based on

1997 MOART-2 run

cloud screening

1996 - 2002

GOME annual changes in tropospheric NO

2

A. Richter et al.,

Increase

in

tropospheric

nitrogen

dioxide

over

China

observed

from

space

, Nature, 437 2005

Slide35

NO2 Trends: Comparison with bottom up estimatesOverall pattern in emission data base is correctIncrease in China is underestimatedIncrease in India and Middle East is overestimatedDecrease in Europe / US is underestimated

35

GOME and SCIAMACHY

EDGAR v4.2

Hilboll, A., Richter, A., and Burrows, J. P.: Long-term changes of tropospheric NO

2

over megacities derived from multiple satellite instruments

, Atmos. Chem. Phys.

13, 4145-4169, doi:10.5194/acp-13-4145-2013, 2013

Slide36

Until 2011, there was continuous increase in NO2After two years of stagnation, 2014 saw a large decreaseeconomic slow down? Improved technology?Switch in fuels used?Other factors?

Satellite NO

2

Trends above Central Eastern China

Slide37

SCIAMACHYInstrument: 8 channel UV/vis/NIR grating spectrometernadir, limb + occultation measurementsOperational on ENVISAT 8.2003 – 4.2012Spatial resolution (30) 60 x 30 km2Global coverage: 6 daysSpecies: O3, NO2, HCHO, CHOCHO, BrO, IO, SO2, H2O, CH4, CO2, CO

37

scanner modules

telescope

pre-disperser

UV channels 1-2

Vis channels 3-4

NIR channels 5-6

SWIR channels 7-8

www.sciamachy.de

Slide38

Buchwitz et al., ACP, 2007; Schneising et al., ACP, 2008Buchwitz et al., personal communication, 2014SCIAMACHY: CO

2

in the Northern Hemisphere

38

Detection of annual cycle

Detection of year-to-year increase

Detection of spatial variability

Not yet accurate enough for Kyoto monitoring on country level

SCIAMACHY and GOSAT: CO

2

Slide39

OMIInstrument: UV/vis imaging grating spectrometer (push-broom)Operational on Aura since October 2004Spatial resolution: up to 13 x 24 km2Global coverage: 1 daySpecies: O3, NO2, HCHO, CHOCHO, BrO, SO2

39

www.

knmi

.nl/

omi

/

Slide40

OMI: SO2 columnsSO2 signals from volcanoes in Ecuador and Columbia Clear signature of Peruvian copper smeltersVery large sources of local pollutionEffect of (temporary) shut down and (permanent) implementation of emission reductions (H2SO4 production) can be monitored

40

Carn, S. A., et al., t (2007), Sulfur dioxide emissions from Peruvian copper smelters detected by the Ozone Monitoring Instrument, Geophys. Res. Lett., 34, L09801, doi:10.1029/2006GL029020.

9.2004 – 6.2005

Slide41

Summary and ConclusionsSatellite observations of tropospheric composition in the UV/vis and NIR provide consistent global datasets for many species including major air pollutants such as O3, CO, NO2, and HCHOThe measurements are averaged horizontally and vertically which makes them difficult to compare to point measurementsRemote sensing in an indirect method that necessitates use of a priori information in the data retrieval which has an impact on the resultsVisible and NIR measurements provide good sensitivity to the boundary layerIn spite of the relative large uncertainties involved in satellite remote sensing , they provide a unique source of information on the composition of the troposphere

41

Slide42

What is the future of UV/vis satellite measurements of tropospheric trace gases?Satellite measurements will be improved byBetter spatial resolution (Sentinel 5P, Sentinel 5, CARBONSAT)Better temporal resolution (geostationary observations Sentinel 4)Better coverage of species and vertical resolution (extension of the wavelengths covered (from UV to IR)High vertical resolution (active systems)The usefulness of satellite data will be improved by better integration with other measurementsSatellite data will be strongly integrated in atmospheric models

42

Slide43

Active measurements: CALIOP aerosol

43

http://www-calipso.larc.nasa.gov/

Slide44

Thank you for your attention

and

questions

please!

http://www.animationlibrary.com/animation/25494/Alarm_jumps/

44

www.doas-bremen.de