page1 sciamachy calibration review – estec – 9 - 13 sept 2002 sciamachy tangent height...
TRANSCRIPT
page1SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
SCIAMACHY Tangent Height Verification
H.Bovensmann
University of BremenInstitute of Remote Sensing
http://www.sciamachy.de
page2SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
Content
Expectations from Instrument On-Ground Verification
General Approach Limb tangent height verification Occultation tangent height verification Summary/Future Work
page3SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
On-Ground Verification
IFOV 0.045° (2.8 km) Absolute pointing knowledge error 0.079° (4.6
km) Relative pointing knowledge error 0.027° (1.6
km) Cycle step size error 0.003° (170 m)
page4SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
General Approach
Use features in the backscattered or transmitted radiance with well known height dependence Absorption of O2 and CO2
Scattering and absorption characteristics at 305 nm around 45 km (so-caled „knee“)
Retrieve tangent height from this „features“ Compare retrieved tangent height to data from
L0/L1
page5SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
Limb tangent height verification
Idea: In the UV the backscattered radiance has a maximum („Knee“) in the upper stratosphere determined by dominating Rayleigh scattering (number density) above and (optical thick) Ozone absorption below the maximum (Janz et al. 1996)
UV-Knee, Sioris et al. TRUE (modified Knee), J. Kaiser et al.
page6SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
0
20
40
60
80
100
0.1 1 10 100 1000
= 260 nm = 270 nm = 280 nm = 290 nm = 300 nm
Uncalibrated limb radiance [BU]
Ta
ng
en
t he
igh
t [km
]
UV limb radiance profiles
optical thin
optical thick
„knee“
page7SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
UV-Knee (Sioris et al.)
find the knee heights for each spectral pixel around 305 nm on the original TH grid
find the knee heights simulated by a model (single scattering, O3 from model atmosphere)
fit a curve to the wavelength (x-axis) vs knee z (y-axis) for the simulation and especially for the vertically undersampled observations (results have also been obtained for knee height vs pointing angle).
compare THs given by the polynomial expressions for observation and model.
Accuracy of better than 220 m is deduced for Orbit 1230
page8SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
UV-Knee (Sioris et al.)
page9SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
Tangent Height Retrieval by UV-B Exploitation (TRUE) (J. Kaiser et al.)
Generalised Knee Method (Janz et al. 1996) normalised model spectra (UV-B) for all tangent
heights are fitted simultaneously to the measured spectrum by varying: tangent height (mean and step size), O3 concentration profile pressure profile
Recommended fit interval 289 – 310 nm (ch. 1b)
page12SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
theoretical retrieval precision reaches 100 m for tangent heights 33–54 km
For 1.6 km step size measurements the following results were observed: mean tangent height error: 0.119 km tangent height step: 1.657 km tangent height step error: 0.012 km
Method is also sufficient to investigate parallelism of limb azimuth scan
Results consistent with Sioris et al. (220 m accuracy)
Report on methodology and first results is available
TRUE Results
page14SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
Occultation tangent height verification
O2 B band absorption ch. 4 (690 nm) 10 – 65 km
CO2 absorption ch. 6 (1560 nm) 10 – 20 km L0 data Solar Occultations from orbit 514, 947 L1 data Solar Occultations from orbit 2056
Systematic shift of approx. 2 km observed w.r.t. L1 data
page15SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
Occultation
page16SCIAMACHY Calibration Review – ESTEC – 9 - 13 Sept 2002
Summary
Limb tangent height accurate within 100 – 200 m Limb pointing performance seems to be an order
of magnitude better than predicted from on-ground performance values (SCIA & ENVISAT)
Occultation tangent height shows a systematic deviation of approx. 2 km
Future Work: Systematic analysis of orbit variations in tangent height Data Analysis w.r.t. parallelism azimuth scan of limb
measurements based on high data rate data with consolidated β-states (Tint = 0.375 s down to 290 nm)
Documentation of results