Transcript - ESA Earth Online

Data processing and quality control activities in support to the
GOMOS mission
F. Niro(1), D. Del Cavallo(1), G. Barrot
(2),
T. Fehr(3)
(1) Serco SpA, [email protected], [email protected] ; (2) ACRI, [email protected] ; (3) ESA/ESRIN, [email protected]
Abstract
DPQC activities
The objective of this paper is to highlight the importance of the satellite instrument and products monitoring and to
show the implication of this task for the scientific user community.
The DPQC is a consortium of specialized companies collecting
engineering and scientific expertise all over Europe. The DPQC role
is to provide on behalf of ESA the service of instrument and product
monitoring and processor implementation for the ENVISAT and ERS
mission. While performing these tasks the DPQC interacts with ESA
as well as with the industry and the science user community.
Whenever an anomaly is detected during the quality check an
investigation is started by involving all the necessary expertise. The
outcome of the investigation is a remedy that could imply a
workaround solution or a re-processing instance; in some cases the
solution could drive the evolution of the processor baseline and/or
the review of the calibration and validation plan.
The final outputs of the DPQC service are the daily and monthly
reports that are delivered to the science user community.
Furthermore the DPQC contributes to workshop and meetings in
support to the mission. A flowchart showing the activities
performed by the DPQC and the main entities involved in the
anomaly handling process is shown on fig.1.
The quality of the satellite data can be variable due to several sources
of instrument and products degradation due to:
in
the
satellite
 The Ground Segment (e.g.: anomaly in activation of ADF,
operational processor anomalies)
The quick detection of all instrument and product anomalies is crucial for instrument safety, as well as for
optimizing the quality of the products before delivery to the users. Furthermore the investigation on the detected
anomaly could drive the development of the operational processor and could help to optimize the instrument
operations and the calibration plan. In fact the quality control is a dynamic process that evolves in order to satisfy
the new requirements coming out during the mission.
In this paper we will address the quality control measures focusing on the GOMOS mission. The mission and
processing status and the monitoring baseline and evolution will be outlined; furthermore the implications of these
activities for the scientific user community will be underlined.
Mission and processing status
Tab. 1 – GOMOS processing configuration at PDS.
The GOMOS mission was interrupted during Jan-Aug
2005 due to the VCCS anomaly. The anomaly consisted in
the failure of the telescope elevation drive.
A workaround solution was found and the mission was
resumed with a horizontal azimuth range reduced to 25°
in order to limit the anomaly occurrence. The main impact
of this anomaly on the GOMOS mission is the reduction
of the data availability: from 35 occultations per orbit to
22 occultations.
No impact on the data quality was reported due to this
anomaly. This anomaly was not detected anymore during
nominal operations since Jan 2006.
DPM
TDS
IODD
GOPR
IPF
6.0
5.5
6.6
6.6
5.1
(3H)
5.4f
5.4c
4.02
4.02
6.1
6.0
7.1
7.1
6.1
(3I)
6.0a
6.0b
6.1
6.0
7.1
7.1
6.1
(3I)
6.0b
6.0c
6.3
6.2
Fig. 1 – Flowchart showing the DPQC anomaly handling process.
Monitoring scheme
7.2c
6.1
(3I)
6.0c
6.0f
6.0d
6.0h
Comment
IPF operational before 8 Aug 2006
-
1st Reprocessing with GOPR
-
Most recent release, includes SCR_1,
SCR_2 & SCR_3. L2 did not pass
verification: minor modifications required
5.0
2nd Reprocessing with GOPR, includes
resolution of all SCR/SPR (except
SPR_19 and SPR_20)
5.0
IPF switch on 08-AUG-2006
Current operational IPF
8th
On the processing side the latest operational IPF was 5.0 that was switched on
Aug 2006 for NRT production and
at FincoPAC for consolidated products, note that Off-line products are available to the users on D-PAC ftp server.
The processing configuration for GOMOS is shown in Tab. 1.
Monitoring results
Daily monitoring
QUADAS reports are inspected daily in order to detect for anomalies on the Flight or Ground Segment.
The level 0 daily monitoring provides a first check of the operational processing performance, since it allows
detection of anomalies in the acquisition or transmission of the instrument source packets. It is important to stress
that the quality of the higher level of production (L1 and L2) depends already from the consistency of the level 0
data. Furthermore the level 0 monitoring it is important for checking instrument behavior, especially device
temperatures.
The monitoring of L1b products allows for checking the temperatures of spectrometers and photometers, the dark
signal per products and the tangent altitude at which the star is lost.
The most interesting analysis presented in the level 1 daily report is the monitoring of the SATU-Y variation along
the mission, in fact since Apr 2006 strong oscillation were observed in this parameter and this problem is still under
investigation. The Star Acquisition and Tracking Unit (SATU) is the CCD that tracks the stars during occultations,
the SATU monitoring consists in statistical angular variation of the SATU data above the atmosphere. Statistics (in
microradians) above 105 km are computed for every occultation, giving four values per occultation: one in the 'X'
direction and one in the 'Y' direction for dark and bright limbs. A mean value per day in every direction and limb is
calculated and monitored in order to assess instrument performance in terms of star pointing. The long term trend
of the SATU 'X' and 'Y' standard deviations should be constant during the whole mission. The actual situation is
shown in Fig. 2. The anomaly started on Apr 2006 is clearly visible in the plot. The reason of the problem is still not
fully clear, attempt were made to reduce it by changing the starting elevation. However since Nov 2006 the starting
elevation was set back to 130km and the anomaly seems to remain stable around a value of 2 microrad.
Tab. 2 – Main quality indicators for GOMOS products.
Product
The following GOMOS products are operationally generated
on the Ground Segment:
 Level 0: Instrument source packets and telemetry data
 Level 1b: Localized calibrated transmissions and
photometer fluxes, one product for each occultation.
 Level 2: Vertical and line density profiles of O3, O2, NO2, O2,
H2O, OClO, air density, aerosols, temperature, turbulence
LEVEL 1b
The quality check provided by the DPQC is extended to all
product levels. The main quality indicators in the GOMOS
products are in tab. 2. The QUADAS tool is used to extract
the quality information parameters from the products and to
generate a wide variety of plots and tables. Final outputs of
QUADAS are html reports but also a database that can be
used for long term analysis.
An important result of the monthly report is the
monitoring of the tangent altitude at which the star is
lost. It is an indicator of the pointing performance
although it is to be considered that star tracking is also
lost due to the presence of clouds and hence not only
due to deficiencies in the pointing performance.
Therefore, only the detection of any systematic longterm trend is the main purpose of this monitoring.
Fig. 4 shows the monthly average of tangent altitudes
in which the stars are lost since Dec 2002. A part from
the peak of Jun-Aug 2005, when testing measurements
were performed, this parameter remains stable along
the mission. This analysis demonstrates that SATU-Y
anomaly is not impact the tangent altitude lost.
Dataset
SPH
Fields
Comment
star_mag
Star magnitude
star_temp
Star temperature
bright_limb
0 = Dark
LEVEL 2
1 = Bright
0 = Full Dark
1 = Bright
2 = Twilight
3 = Straylight
4 = Twi.+Stray
SUMMARY
QUALITY
obs_ill_cond
AUXILIARY DATA
pcd
PCD at measurement level
pcd_sp
Product Confidence Data at pixel level
for the spectrometers
pcd_fp
Product Confidence Data at pixel level
for the photometers
LOCAL SPECIES
DENSITY
pcd
PCD indicated if profile point is valid
AEROSOLS
pcd
Product Confidence Data indicated if
profile point is valid
TRANSMISSION
Monthly average of tangent altitude loss
45
Tg altitude (Km)
 The Space Segment (e.g.: anomaly
communication, instrument ice contamination)
40
Mean tg altitude loss dark
35
Mean tg altitude loss bright
30
Std tg altitude loss dark
Std of tg altitude loss bright
25
20
15
10
5
0
Jul-02 Dec-02 Jun-03 Dec-03 Jun-04 Dec-04 Jun-05 Dec-05 Jun-06 Dec-06 Jun-07
Date
Fig.4-Monthly mean tangent altitude (and STD) at which the star is lost.
Another important result of the MR is the monitoring
of the Product Quality information stored in level 1b
products (i.e. number of invalid measurements,
number of measurements containing saturated
samples,
number
of
measurements
with
demodulation flag set…).
The output of this analysis is plotted in the world
maps of fig. 5 in terms of % of cosmic ray hits per
profile, % of datation errors per profile, % of star
falling outside the central band per profile and % of
saturation errors per profile.
Looking at fig. 5 the high percentage of cosmic ray
hit can be clearly observed when the satellite
crossed the South Atlantic Anomaly (SAA) region.
Also the percentage of saturation errors per profile
shows an increase over the SAA zone.
Fig. 5-L1b quality monitoring with respect to geo-location of ENVISAT
Fig. 2-Trend of SATU NEA std deviation since April 2006 and starting altitude of occultations.
The monitoring of the level 2
products consists in the check of the
processing
consistency,
format
integrity and scientific reliability of
the atmospheric data.
One interesting section of the L2
daily report is presented in Fig.3
showing the Ozone profiles, from left
to right the first plot shows all values
including outliers, the second is a
zoom on the stratospheric peak and
the last on the mesospheric peak.
Similar figures are reported for
different standard deviations.
Another example of monthly L2 monitoring is
the plot presented in Fig. 6, it is the average of
the Ozone values during the reporting month
in a grid of 0.5 degrees in latitude per 1 km in
altitude. Only occultations in dark limb have
been used. Even though there is a reduction
on latitude coverage due to the restricted
azimuth field of view of the instrument, still
some known characteristics can be seen:
 O3 concentrations show a decrease with
latitude near 40 km altitude. In the lower
latitudes O3 is generated by photolysis of O2.
 In the middle stratosphere (25-30 km) O3 is
strongly influenced by transport effects.
Strong transport towards South region is
visible in middle and higher latitudes.
Fig. 3 –Ozone profile
 The lower stratosphere shows an O3
increase with latitude. Highest values can be
found within higher latitude regions due to
downward transport of rich air masses.
Long term analysis
In addition to the daily monitoring which is aimed at the quick detection of instrument and products anomaly a
systematic long term analysis is carried out on a monthly basis and a report is delivered to the user community.
The MR is issued with the support of the QWG, the report contains information about:




Mission status and instrument planning
Processing Status and Configuration on the Ground Segments
Products quality monitoring and long term analysis
Anomaly reporting and investigation
The main results of the GOMOS MR are the long term analysis of the instrument status and performances, the long
term trend of the gain and spectral calibration and the mispointing evolution. The analysis of significant instrument
parameters on a long term basis is important in order to study long term trend or seasonal variation and for detecting
instrument ageing.
Fig. 6 - Average GOMOS O3 profile for March 2007.
Source of information
Data quality disclaimers provide information on known deficiencies in processing, and on transient degradations not yet
compensated
http://envisat.esa.int/dataproducts/availability/disclaimers/
Daily/monthly reports can be accessed at:
http://earth.esa.int/pcs/envisat/gomos/reports/
The unavailability intervals of the GOMOS instrument are available on line
http://envisat.esa.int/instruments/availability/
For any questions on GOMOS, please contact ESA’s EO Helpdesk
[email protected]