Ice contamination on satellite IR sensors: the MIPAS case

1
Ice contamination on satellite IR sensors the
MIPAS case

• F. Niro(1), T. Fehr(2), A. Kleinert(3), H.
  Laur(2), P. Lecomte(2) and G. Perron(4)
• (1) Serco S.p.A., Via Sciadonna, 24, 00044
  Frascati, Italy
• (2) European Space Agency (ESA) - ESRIN, Via
  Galileo Galilei, 00044 Frascati, Italy
• (3) Forschungszentrum Karlsruhe GmbH, Institut
  für Meteorologie und Klimaforschung (IMK), P.O.
  Box 3640, 76021 Karlsruhe, Germany
• (4) ABB Bomem Inc., 585 Blvd. Charest East,
  Québec, G1K 9H4, Canada

2
Contents

• Intro
• MIPAS on ENVISAT instrument overview and status
• Calibration
• MIPAS Calibration strategy and requirements
• Focus on the radiometric calibration the gain
  function
• Ice effects
• The ice effects on MIPAS
• Ice on AATSR and SCIAMACHY
• Summary
• Summary and lessons learned
• Whats next
• The ENVISAT and MIPAS mission extension beyond
  2010

Intro
3
MIPAS on ENVISAT

• MIPAS is a FTS measuring the limb atmospheric
  emission in the mid-IR spectral range, 4.15 µm
  14.5 µm, 685 2410 cm-1
• The INT is a dual slide. The two output ports are
  directed to two sets of four duplicated
  detectors. This allows for redundancy and for
  enhanced radiometric performances
• The detectors and their fore optics are stored in
  the FPS and cooled down to 70K with a pair of
  Stirling-cycle coolers

Intro
4
Optimized Resolution mission

• On March 2004 after an increase of INT velocity
  errors (speed of one or both slides exceeds 20
  of the nominal speed) ? the MIPAS mission was
  suspended and several tests made to find the best
  configuration
• The new scenario started on Jan 2005 with
• Spectral resolution was reduced to 41 of the
  original one (0.0625 cm-1 instead of 0.025 cm-1)
• Vertical and horizontal sampling of the
  atmosphere was increased owing to the shorter
  measurement time
• Duty cycle reduced to about 40 in order to
  reduce INT errors

Intro
5
MIPAS instrument status

• High number of INT velocity errors were still
  observed during 2005 2006
• The errors type was analyzed in details it was
  found that it depends on Temp (beam-splitter),
  friction (bearings) and of initialization
  (start-up)
• Corrective actions were undertaken that allow to
  decrease the number of errors and increase duty
  cycle up to 100 since Dec 2007

Intro
6
MIPAS products and calibration

• The MIPAS operational products are
• Level 1 Calibrated atmospheric emission spectra
• Level 2 atmospheric profiles of p-T and main
  target species (O3, H2O, CH4, N2O, HNO3, NO2)
• The L1 calibration process consists of
• Radiometric calibration The process of assigning
  absolute values in radiance units (W/(cm2 sr
  cm-1)) to the intensity axis (y-axis)
• Spectral calibration The process of assigning
  absolute values in cm-1 to the wavenumber axis
  (x-axis)
• LOS calibration The process of assigning an
  absolute LOS pointing value to a given
  atmospheric spectrum

Calibration
7
MIPAS L1 radiometric calibration

• The radiometric calibration is a crucial step of
  the L1 processing since an error in this
  calibration directly translates into an error in
  the retrieved profiles. The radiometric
  calibration requires
• Deep space (DS) measurements to correct the scene
  for self-emission of the instrument. DS
  measurements are done frequently to account for
  variation of the instrument Temp along the orbit
• Blackbody (BB) measurements followed by an
  equivalent number of DS measurements to calculate
  the radiometric gain function.
• The gain function is calculated once per week,
  this allows to fulfill the requirement of gain
  accuracy (1)
• The radiometric gain G is calculated from the
  measured BB and DS radiances (SBB and SDS), and
  the theoretical BB radiance (LBB)
• The measured radiance (LX) is calibrated using
  this gain function, the observed radiance of the
  scene (SX) and of the offset (Sc) closest in time

Calibration
8
Weekly changes of gain

• Changes in the gain function are caused mainly by
  changes in instrument transmission, due to ice.
  The ice is somewhere in the optical path in the
  focal plane subsystem which is cooled to 70K
• The ice is either on the edges of the entrance
  hole (of the focal plane subsystem) and/or on the
  dichroics (splitting the light to detectors)
  and/or the detector windows
• Often the MLI (multi-layer insulator) may trap
  water from the air on-ground. This trapped water
  evaporate (outgassing) with increase of
  temperature and can deposit in coldest part of
  the instrument with formation of ice layer

Ice effects
Gerakines et al., Astronomy and Astrophysics 296,
810, (1995)
Ice absorbance
Gain changes
9
Gain variation in band A

• Ice accumulates on optics with loss of signal at
  the detector, ice is released after
  decontamination (cooler switch-off)
• The variation of position of ice maximum is due
  to variation of ice layer thickness that can
  introduce other effects (e.g., ice scattering)

Ice effects
10
Rate of gain variation

• The requirement of 1 increase/week is fulfilled
  (dashed line)
• We observe an overall decrease of outgassing
  along the mission.
• We observe the very contaminated period of Jan
  Jun 2005, due to the fact that decontamination
  was not planned during Feb Dec 2004

Ice effects
11
Gain and NESR

• The NESR of the scene is defined as the standard
  deviation of the measured single sweep spectral
  radiance taken over N measurements
• Gain and NESR variations are linearly correlated
  and similarly degraded by ice contamination (loss
  of transmission)

Ice effects
12
Gain and NESR variation

• NESR variation is linearly correlated to gain

Ice effects
13
Ice effects on L2 precision

• Precision is proportional to NESR ? NESR varies
  due to ice contamination, but it is also slightly
  dependent on signal ? higher radiances means
  higher NESR
• Precision (random error due to noise) on VMR
  retrieval is inversely proportional to Temp
  (Planck function) ? Higher Temp ? Stronger signal
  ? Better precision
• The impact of these two factors (ice and
  atmospheric temperature) on the time variation of
  L2 precision is complex (see C. Piccolo and A.
  Dudhia, ACP, 7, 19151923, 2007)
• In general L2 precision degrades proportionally
  to ice contamination
• In case of weak species the L2 precision is
  critically degraded by increasing NESR
• In case of large seasonal variation of
  atmospheric temperature (polar region) ? L2
  precision is more driven by variation of
  temperature
• Furthermore ice contamination impacts directly
  accuracy of profiles
• An error in the gain function of 1 directly
  translates into a systematic error of 1 in the
  calibrated spectra and then in the profiles

Ice effects
14
Ice in other ENVISAT instruments AATSR and
SCIAMACHY

• Ice was also seen on other ENVISAT instruments,
  AATSR and SCIAMACHY
• It may be that outgassing from other parts of
  ENVISAT lead to increased water vapor pressure
  around the satellite, this water vapor may reach
  the MIPAS detector unit

Ice effects
Signal loss on SCIAMACHY channel 8
Courtesy of SOST-IFE
15
Summary and lessons learned

• Water is trapped on-ground in some parts of the
  platform (e.g., MLI) and it evaporates in flight
  (outgassing) forming ice around the coldest parts
  of the ENVISAT satellite, in particular IR (cold)
  sensors such as MIPAS, AATSR and SCIAMACHY IR
  channels
• Ice formation determines loss of signal at the
  detector
• Outgassing is increasing with Temp during hottest
  period of the year
• Outgassing is decreasing along the mission since
  contaminants are progressively removed from the
  instrument
• Periodic decontaminations should be performed, in
  order to avoid that the decrease of
  signal-to-noise ratio impacts products quality
• The most critical part of the mission is the
  first year, when very strong contamination was
  seen in all ENVISAT IR sensors
• Similar issues were also found during operations
  of other IR sensors in different platform (e.g.,
  IASI and ACE)

Summary
16
MIPAS mission extension

• The inclination control will be switched-off
  starting from 2011 in order to minimize the fuel
  consumption. We will loose the repeat orbit track
  away from the equator and a drifting Mean Local
  Solar Time (MLST)
• Since 2011 the altitude will be lowered by 25 km
  and controlled, while the MLST will be left
  drifting until end of the mission (possibly 2014)
• No showstoppers have been found for MIPAS
  (instrument and processing), however some care
  should be taken in order to avoid sun light
  entering the ESU/ASU

Whats next
17

• Thank you for your attention !
• Acknowledgment
• M. Birk (DLR), G. Davies (VEGA), A. Dehn (Serco),
  A. Dudhia (Oxford University)

Questions / Answers