ESA Summer School - Atmospheric Composition Sounding - B.Kerridge ESRIN, Aug’2010 1 Atmospheric...

ESA Summer School - Atmospheric Composition Sounding - B.KerridgeESRIN, Aug’2010 1

Atmospheric Composition Sounding

ESA Summer School

ESRIN

12th August 2010

Lecture 1 by B.Kerridge

Part 1 – Introduction

Part 2 – mm-wave limb sounding

2

OutlineAim of three lectures: overview atmospheric composition sounding from space

Lecture 1

Part 1 - Introduction1. Scientific imperative2. Operational applications3. Evolution of models & assimilation systems4. Challenges imposed by physics

Part 2 - mm-wave limb-sounding1. Principles2. Instrument attributes3. Linear retrieval diagnostics 4. Odin SMR and Aura MLS5. Summary and future advances

ESA Summer School - Atmospheric Composition Sounding - B.KerridgeESRIN, Aug’2010

3

• The atmosphere is constantly changing

ESA Summer School - Atmospheric Composition Sounding - B.KerridgeESRIN, Aug’2010 Introduction

ESA Living Planet Symposium - Atmospheric Composition - B.KerridgeBergen, 27th June 2010 4

5

• Understanding of atmospheric phenomena, including clouds and aerosol, depends upon remote-sensing observations


6

1. Scientific Imperative

• Atmospheric composition is driving climate change

– Primary radiative forcing by trace gases & aerosol– Indirect effects through chemistry and aerosol-cloud– Feedbacks via water vapour, cloud & trace gases


7

Scientific Imperative (contd)

• Remote-sensing plays a vital role

– to investigate processes– for international assessments (IPCC,WMO)

• Satellites: global to regional scales

• Airborne / surface: finer scales; ground-truth

• Troposphere is challenging


8

Scientific objectives for sounding composition from space

1. Processes controlling distributions of key trace gases within height-range important to climate

2. Links to biogenic, pyrogenic and anthropogenic emissions

– Advances will be made in future by resolving structure on finer-scales to understand processes

– Requirements are stringent.


ESA Living Planet Symposium - Atmospheric Composition - B.KerridgeBergen, 27th June 2010 9

Operational Applications for Satellite Composition Data

Ozone Layer & Surface UV─ satellite global observations have 30-year heritage

Composition – Climate Interaction─ satellites observe trace gas profiles, aerosol & cloud which cannot be monitored adequately from surface alone─ decadal, self-consistent data sets → GCOS essential climate variables (ECVs)

Pollution Monitoring & Air Quality Forecasting ─ satellites observe pollutants above and between surface sites

10

3. Evolution of models and assimilation systems

• Models are evolving rapidly• Increased computer power enables:

– More complete atmospheric physics, dynamics & chemistry – Integration of troposphere and stratosphere– Couplings with land and ocean for climate-composition interaction

• Assimilation systems in development to exploit satellite data in

addition to surface level observations, analogous to NWP• Inverse modelling to quantify surface emissions• Model resolution to increase in coming decade• Processes that were sub-grid scale will be resolved (eg convection)

→ To critically test future models, satellite observations of increased resolution will be needed.


ESA Living Planet Symposium - Atmospheric Composition - B.KerridgeBergen, 27th June 2010 11http://daac.gsfc.nasa.gov/fieldexp/TOGA/cls.shtml

12

– Spectral line interference ← pressure broadening– Water vapour & other continua – Clouds– Discriminating surface ← especially critical for aerosol– Seeing through higher layers ← nb tropospheric O3

– Vertical resolution in tropopause region and near-surface layer• where height leverage from T profile is minimum

– The challenge is: • to achieve significantly higher vertical resolution, horizontal

resolution and spatio-temporal sampling• while preserving/improving accuracy

4. Challenges imposed by physics


13

Part 2 – mm-wave limb-sounding

• Profiling mid/upper troposphere & lower stratosphere– where climate sensitivity is greatest

ESA Summer School - Atmospheric Composition Sounding: - B.KerridgeESRIN, Aug’2010 mm-wave limb sounding

14

1) Principles of mm-wave sounding

1. In microwave – sub-millimetre-wave region

> 100m (0.1mm), < 100 cm-1, f < 3THz

most molecular transitions are pure rotational• Rotational levels closely-spaced in energy ( ~1 cm-1)

• Collisions with N2 & O2 maintain Boltzmann population distributions of rotational levels up to the thermosphere

• TR → Tk – local kinetic temperature

• J(,TR) → B(,TK) – local thermodynamic equilibrium (LTE)

2. Departure from LTE can be significant at mid-IR and shorter ’s─ radiative and photochemical processes can affect populations of vibrational

levels ( ~1,000 cm-1) and electronic levels ( ~10,000 cm-1) in stratosphere and above

─ have to be modelled in some cases, even if targeting low atmosphere


15

Radiative Transfer Equation at Long Wavelengths

,,0,0, dzzvdz

dzvBvvBvR

oo

o

Planck

Function

(Assuming boundary emissivity =1)

1/exp

2,

32

zkThcv

vhczvB

1/ kThcv

zkTcvzvB 22,

If (Rayleigh-Jeans)


16

Radiative Transfer Equation at Long Wavelengths continued

Planck Function linearly dependent on T(z) cf IR wavelengths

“Brightness Temperature” defined as:

dzzvdz

dzTvTvT

kcvvRvToo

o

B

B

,0,0

2/ 2

– Simplest form of atmospheric radiative transfer equation, except for

direct-sun absorption.– Applies unless surface reflectivity (nadir) or cloud scattering are non-zero– For limb geometry, boundary is “cold space” .


17

Atmospheric transmittance spectra for limb-geometry(H2O, O2 & O3 only) 0 – 1000 GHz

12km

10km

8km

6km

Wings of strong H2O lines control spectralcurvature & penetration depth in limb-views

Troposphere seen in limb-views <380GHz

Frequency (GHz)

Tra

nsm

itta

nce


Wavelength Dependence of Cirrus Extinction

─ Mm-wave affected only by scattering from large ice particles (> 80m)→To observe smaller cirrus size components would require frequencies up to ~THz

─ Thin tropical tropopause cirrus, PSCs and aerosols: Re < 10 m→ These particulates transparent in mm-wave

MARSCHALS

IR sounders egMIPAS & IASI

1mm


2) Instrument Attributes

─ Heterodyne (coherent) detection:

1. Signal from atmosphere mixed with local oscillator (LO) signal─ Down-conversion to intermediate frequency (IF) band for amplification─ Either upper & lower frequency bands superposed (DSB) or one band filtered out (SSB)

or side-bands separated

2. Spectral resolution intrinsically high (bandwidth a more limiting factor)─ Doppler(1000m) = 0.01 x Doppler(10m) → Lines p-broadened throughout strat. and fully-resolvable →height leverage

3. NEBT (K) ~ Tsys (K) / √ [ int (s) x f (Hz) ]

4. A ~ 2 ← Throughput determined by wavelength

─ Long wavelengths: diffraction limited optics

─ From polar orbit at 800km, vertical half-power beamwidth (HPBW) of 2km at tangent-point requires ~1.6m antenna at 1mm

─ For 1mm, HPBW proportionally larger for same antenna size.


20

Spectra Simulated for MLS bands at 190 GHz (R2) and 240 GHz (R3) bands

H2O

HNO3

CO

O3 O3O3 O3 O3 O3

HNO3


21

H2O Weighting Functions for MLS


22

O3 Weighting Functions for MLS


23

CO Weighting Functions for MLS


3) Linear Retrieval Diagnostics


Influence of antenna width & oversamplingon profile retrieval precision & resolution

• Retrieval of H2O, O3, CO & HNO3 profiles simulated for idealized case:─ MLS 190 & 240 GHz bands; contiguous, f =100MHz─ MLS Tsys(DSB) 1000K; scan range & duration fixed ─ 2km retrieval level spacing; √So(i,i) = 100% (diagonals)─ Antenna: nominal HPBWs (4.2 or 3.2km) or pencil beam─ Limb-view spacing: 300m or 2km

• Retrieval linear diagnostics examined:─ Averaging kernels, AK FWHM and √ [ Sx(i,i) / So(i,i)]


26

Averaging Kernels for MLS

Averaging kernels indicate:– vertical resolution of retrieval– improvement on prior knowledge

HPBW: ~4km (190 GHz) , ~3km (230GHz)

→ Vertical resolution <HPBW achievable for H2O and O3

S tangent-height spacing 2kmS p-broadening


Influence of antenna beam width and limb view spacing on retrieval: O3 & H2O ESD & AK FWHM

antennaconvolved

antennaconvolved

not convolved

not convolved

300m spacing

300m spacing

R1 Antenna:HPBW: 4.2km



Influence of antenna beam width and limb view spacing on retrieval: HNO3 & CO ESD & AK FWHM

antenna not convolved

antenna not convolved

R1 AntennaHPBW: 4.2km



Antenna pattern width & oversampling

─ Vertical resolution <HPBW achievable from spectral lineshape and oversampling─ Provided radiometric sensitivity sufficiently high ─ Also assuming: perfect knowledge of antenna-pattern & tangent-point spacings

─ For real antenna and retrieval level spacing of 2km, additional benefit from reducing limb-view spacing from 2km to 300m is limited.


30

Key Instrumental Uncertainties

• Pressure of a reference limb-view in scan & T profile can be retrieved accurately

• In addition to NEBT, precision & accuracy of constituent retrieval depend on knowledge of:

─ Vertical spacings of limb-views

─ Vertical shape of antenna pattern (for convolution of pencil beams)─ Critical for upper troposphere: H2O vertical gradient; clouds lower down

─ Beam efficiency and loss─ BB calibration target cannot be in front of antenna

─ Frequency-dependent responses of upper & lower side-bands ─ Particularly for “double side-band” receivers

─ Spectral baseline─ Especially structure on scales similar to atmospheric emission features


4) Odin SMR & Aura MLS

─ Odin Sub-Millimetre Radiometer (SMR) ─ Aura Microwave Limb Sounder (MLS) ─ Satellite limb-emission sounders launched 2002/3


ODIN SMR Technical Specifications

The Instrument:

•Antenna size: 1.1 m •Beam size at 119/550 GHz: 9.5'/ 2.1'(126") •Main beam efficiency 90% •Pointing uncertainty <10" (rms)

•Submm tuning range: 486–504, 541–581 GHz•Submm Tsys(SSB): 3300 K ← Cooling to 80K

•HEMT:118.75 GHz; Tsys (SSB): 600 K

•AOS b’width / res.:1100 MHz / 1 MHz•DAC b’width /res:100–800 MHz / 0.25–2 MHz

The satellite:

Orbit: sun-synchronous dawn-dusk, polar orbit, altitude 600 km

Platform: 3-axis stabilized (reaction wheels, star sensors, gyros)

Mass: 250 kg (bus 170 kg, instruments 80 kg)

Size: Height 2m, width 3.8m (incl. solar panels)


Example Single Side-Band Limb Spectra from SMR

Enhancedby PSCs

Descent in vortex of air low in N2O

– Odin SMR sees through PSCs to observe trace gases in Antarctic ozone hole


Odin SMR Distributions in S.Hem. Lower Strat. 19th Sept – 5th Oct 2002

500K surface ~ 20km

Vortex centred:─ N2O low – desc. ─ NOy low – denit.─ ClO high – act.→ O3 hole

Warmingsplits vortex

Vortex reforms

ClO converted back to Cl reservoirs


Aura MLS in the A-Train

MLS views forwards and limb-scansin the orbit plane


Aura MLS Technical Specifications

See intotroposphere


Example Double Side-Band Limb Spectra from MLS

–By observing in higher frequency bands than Odin SMR, MLS detects two additional species important to stratospheric ozone chemistry: HCl (~625 GHz) and OH (~2.5 THz)


Tomographic Limb Sounding

185km

Traditional: – concentric, homogeneous layers – 1-D profile retrieved from single-scan

Tomographic: – 2-D structure in radiative transfer model– 2-D field retrieved from multiple-scans


Principles of Tomographic Limb-Sounding

─ 2-D RTM instead of spherically-symmetric atmosphere ─ State-vector: 2-D grid instead of 1-D profile─ Measurement-vector: set of limb-scans which are inverted

simultaneously─ Limb-scan spacing along-track fine enough to oversample retrieval grid

or apply regularisation─ Given air volume viewed from many different directions → tomography

─ Demonstrated operationally by Aura MLS


Practical Considerations

– Iterative solution to Optimal Estimation equation:

– Sa is a priori covariance matrix of x – K is weighting function matrix w.r.t. x– Sy is measurement error covariance matrix

– Current memory limitations preclude storage of matrices of dimension Ny

– Provided Sy is diagonal, matrices such as KTSy-1K can be accumulated sequentially

on limb-view by limb-view basis

– Since Nx is also large, further matrix manipulation required to make the problem computationally viable.


41

MLS maps – 215hPa 12-18th Jun’05

Jonathan Jiang, MLS team, JPL

• In addition to profiling many gases in the stratosphere, MLS can also probe H2O, O3 & CO in the upper trop. • CO and O3 higher than in GEOS4-CHEM

– agreement improved in more recent processing• HNO3 also now extended below tropopause


Seasonal Average Retrieval of Ice Water from Aura MLS & Odin SMR

MLS SMR


43

Pyroconvective Plumes

Calipso 10th Feb’2009Courtesy NASA Langley

MLS COCourtesy H.Pumphrey, U.Edinburgh

Plumes from Australian fires Feb’09 observed to enter lower stratosphere by Calipso & MLS

ppbv

Day ‘09


44

ECMWF analysis – sonde RMSE comparison: Aura MLS O3 NRT data

Jul-Aug 2008

CTRLMLS

60-90S30-60S

60-90N 30-60N 30S-30N

Courtesy, R.Dragani (ECMWF)

Impact of Limb Emission Data in Assimilation

S Assimilation of MIPAS NRT radiances under assessment


South Pole ozone profiles from GEMS reanalysis

Ozone profile data important for assimilation

Oct 2003 Oct 2004 Oct 2005 Oct 2006 Oct 2007

MIPAS MLS MLS

46

5) Advances in limb-emission sounding

• Profiling mid/upper troposphere & lower stratosphere:

– Improve vertical resolution, horizontal resolution & coverage– Improve accuracy eg handling of cloud→ Tomography (2-D → 3-D)

• Foreseen technology developments:

– mm-wave: system noise limited → scope for improved technology• Receiver array: rugged planar mixers with integrated IF amps• Side-band separating mixers → both side-bands are SSB • Back-end spectrometer with broad bandwidth (eg DAC)• SIS mixer (4K): Tsys lower than Schottky-diode mixer

– Optical: detectors photon noise limited in mid-ir • Fixed staring 2-D detector array, coupled to FTIR


ESA Living Planet Symposium - PREMIER - B.KerridgeBergen, 30th June 2010 47

MARSCHALS mm-wave limb sounder

• ESA airborne simulator for space sensor─ to demonstrate new observing capabilities

• Design optimised for upper troposphere─ Bands at 300, 325 & 350GHz; SSB─ 12GHz bandwidth, 200MHz resolution

• Campaigns in tropics & Arctic

• Observation of H2O & O3 through sub-tropical cirrus

MARSCHALS Observations

Frequency (GHz)

Cloud opaque in IR & near-IR limb-views

mm-wave limb spectra co-located 0.75m limb imager

BT (K)


5. Summary & Future Advances1. Millimetre-wave limb-sounding is a powerful tool for profiling

stratospheric trace gas distributions from space:─ Observations are insensitive to aerosol and PSCs and to non-thermal

emission processes

2. Capability now extended to profile trace gas distributions in the upper troposphere─ A major advantage is that most cirrus clouds are fully/semi transparent at

mm-wavelengths in limb-geometry─ Ice water distribution in upper troposphere retrieved by Odin SMR /Aura

MLS from size component ( >80m) which is seen at mm-wave.

3. Advances foreseen for future missions: ─ Optimisation for sounding the upper troposphere ─ Significantly increase vertical & horizontal resolution & accuracy

→ Observe finer-scale structure in distributions of key trace gases (eg H2O, O3, CO, HNO3)

→ Investigate processes controlling composition in the height-range of particular importance to climate


Date post:	16-Jan-2016
Category:	Documents
Upload:	clifton-hines
View:	215 times
Download:	0 times

ESA Summer School - Atmospheric Composition Sounding - B.Kerridge ESRIN, Aug’2010 1 Atmospheric...

Documents