ESA Earth Home Missions Data Products Resources Applications
    24-Jul-2014
EO Data Access
How to Apply
How to Access
Index
  Acknowledgements and contact
List of figures and tables
List of tables
 Additional information on the GOMOS measurements
Periods of data unavailabilities
Evolution of the IPF
Glossary
List of abbreviations and acronyms
Product types and structure
Other sources of information
GOMOS-related ESA web pages
Publications (peer-reviewed publications and proceedings of conferences)
Secondary products
Scintillation and turbulence
Aerosols and PSC
Product validation
NO2 and NO3 measurements
O3 measurements
CAL/VAL activities
Retrieval and processing issues
Assimilation of GOMOS products
Mesosphere
Specific events
GOMOS-related theses
Other technical reports by members of GOMOS SAG, ESL and QWG
Level2 processing
How to
FAQ
Access tools
GOMOS products toolbox
EnviView
Data selection
PCD summary in the Level2 products
Obliquity
Star properties
Data availability
Presentation of the GOMOS products
Description of the products
Product content
Product structure
Time availability
Data size
Presentation of the instrument and the measurements
Scientific achievements
Validation results
Measurement characteristics
Accuracy
Occultation obliquity
Geographical and time coverage
Star characteristics
Mission planning
Modified mission scenario since August 2005
Instrument description and measurement principle
Calibration phase and monitoring activities
Measurement technique
GOMOS Product Handbook
Content of the products
Services
Site Map
Frequently asked questions
Glossary
Credits
Terms of use
Contact us
Search


 
 
 


5.1.3 Content of the products

5.1.3.1 General common fields to all products

SPH descriptor

ASCII string describing the product.

Start time

start time of the measurements included in the MDSR. Expressed in Julian Date 2000.

Attachment flag

flag indicating if MDSRs are attached to the current ADSR
  1: no corresponding MDSRs
  0: there are corresponding MSDRs

In the case of GOMOS, the attachment flag is always set to 0, indicating that a MDSR is attached to the ADSR.

Data Quality indicator

indicator of the quality of the data included in the MDSR
  -1: blank MDSR
  0: data are available and valid

In the case of GOMOS, the attachment flag is always set to 0, indicating that the MDSR is not empty (as there is no empty MDSR in the GOMOS products).

Creation time

creation time of the data set record. Usually inserted at the beginning of the DSRs corresponding to calibration data. Expressed in Julian Date 2000 (transport format).

Spare

dummy bytes.

 

5.1.3.2 Level 1b product

5.1.3.2.1 Definitions related to the SPH and the DS

Definitions related to the SPH

Start time of the occultation

Starting time of the first acquisition of the occultation.

Time of the end of the occultation

Ending time of the last acquisition of the occultation.

Latitude and longitude of the tangent point at the start time

Location of the tangent point at the starting time of the occultation. Positive latitudes are North; positive longitudes are East. Computed during the geolocation process. Provided for the reference wavelength, which is equal to 500 nm.

Latitude and longitude of the tangent point at the stop time

Location of the tangent point at the ending time of the occultation. Positive latitudes are North; positive longitudes are East. Computed during the geolocation process. Provided for the reference wavelength, which is equal to 500 nm.

Occultation duration

Duration of the occultation expressed in seconds. Computed during the datation process.

Sampling duration

Nominal value is 0.5 second.

Number of measurements

Number of acquisitions performed during the occultation. This number is computed when the GOMOS packets are read from the Level 0 product.

Status of the instrument

The Redundancy Definition vector and the Instrument Configuration fields are used to determine in which mode the GOMOS instrument was during the measurements. These two fields are read in the Source Data Field Header of the first valid packet of the occultation.

Status is set to 0 if the instrument is in extended mode i.e. if:

  • Bit 12 of the Redundancy Definition vector set to 1
  • Bit 13 of the Redundancy Definition vector set to 0
  • Bit 14 and 15 of the Instrument Configuration set to 1
  • Bit 4 and 5 of the Instrument Configuration set to 0

and status is set to 1 in another configuration

Occultation number in the orbit

This information is provided by the Source Data Packet extraction processing.

Star name and associated information

These information are read from the star catalogue auxiliary product once the star identifier of the current occultation has been read in the GOMOS packets.

Star identifier in the star catalogue

Star identifier read in the Source Data Field Header of the first valid packet of the occultation.

Star direction in the quasi-true of date frame (Cartesian)

Star direction computed in the geolocation processing.

Dark/bright limb flag

Flag read in the GOMOS packets. A value of 0 means that the electronic gain settings of the instrument has been set to optimise the measurement of the star signal (this is usually called dark limb settings), while a value of 1 means that the gain settings has been set in order to perform the observation assuming some limb signal (this is usually called bright limb settings). Note that this value is defined in the mission planning and, although it is generally in line with the actual illumination conditions, it may not reflect the actual illumination conditions. The real illumination condition is given by the illumination condition PCD and is computed from the sun zenith angle at both the instrument and the tangent point locations.

 

Definitions related to GADS

Summary quality

The meaning of the possible flag values stored in the Summary Quality GADS is detailed in section 5

Number of points of the spectra

The programmable number of transmitted columns per CCD array. The values are read in the GOMOS packets.

In the Level 1b product, the data coming from the 4 CCDs are joined together in a single vector. This is true for the transmission and covariance spectra but also for the reference star spectrum, the limb spectra and the wavelength assignment vectors. This parameter must be used to distinguish between the different CCD arrays inside these global vectors. For instance, if this parameter is equal to the current nominal values: [450, 966, 420, 500], it means that in each global vector (e.g. transmission) the 450 first samples belong to the UV CCD (SPA1), the following 966 belong to the VIS CCD (SPA2), the 420 next to the IR1 CCD (SPB1) and the last 500 to the IR2 CCD (SPB2).

The total number of transmitted columns is currently equal to 2336.

Number of photometer output data per measurement

This is the number of fast photometer output data per spectrometer measurement. Nominal value is 500. This number will never change during the instrument life.

Number of SATU output data per measurement

This is the number of SATU output data per spectrometer measurement. Nominal value is 50. This number will never change during the instrument life.

Photometer central wavelength

This is the wavelength corresponding roughly to the middle of the spectral range of each photometer. The nominal values are 499.5 nm for the blue photometer and 672.0 nm for the red one.

Time shift for ray tracing / geolocation

This time indicates the time shift between what is called 'beginning of measurement' and 'half measurement' in the ray tracing processing. Geolocation results are provided at these two times for each measurement.

Spectrometer effective sampling time

Actual duration of one spectrometer measurement. Nominal value is 0.4999639 second. This is a constant value.

Reference wavelength for the ray tracing

All the geolocation information is provided for this reference wavelength. For example the localisation of the ray nodes and of the tangent point node are provided for this wavelength although at one given time inside the atmosphere, and due to chromatic refraction effects, each wavelength is looking at a different altitude. It is also the reference wavelength for the aerosol products. It is equal to 500 nm.

Radiometric sensitivity curve (background)

This curve has been built from the radiometric sensitivity of the instrument measured during its on-ground characterisation and updated in-flight. It allows the conversion of the background spectra expressed in electrons in the product into physical units (ph/s/cm2/nm/sr). It is given as a LUT of conversion factors. A linear interpolation of the conversion factor is needed to use this curve for any sample of the background spectra.

Radiometric sensitivity curve (star)

This curve has been built from the radiometric sensitivity of the instrument measured during its on-ground characterisation and updated in-flight. It allows the conversion of the star spectra expressed in electrons in the product into physical units (ph/s/cm2/nm). It is given as a LUT of conversion factors. A linear interpolation of the conversion factor is needed to use this curve for any sample of the star spectrum.

Thermistor temperature (SP and FP)

Thermistor temperature read from the GOMOS packets and converted into degrees using a conversion LUT calibrated on-ground. This temperature is used to adjust the dark charge map to the actual temperature. The actual thermistor temperature are transmitted once per occultation, in the first packet. The coding accuracy of the thermistor temperature is roughly 0.4°.

Spectrometer dark charge at band level used for the dark charge correction

Dark charge at band level for each of the 3 bands (upper, central and lower) and for each of the transmitted CCD column.

Mean dark charge for the spectrometers

The mean dark charge level over the CCD array, expressed in electrons. This mean dark charge is assumed to be constant during all the occultation. It is provided in this order: SPA1 upper, target and lower bands, SPA2 upper, target and lower bands, then SPB1 and SPB2.

Mean dark charge for the photometers

The mean dark charge level over the CCD array, expressed in electrons. This mean dark charge is assumed to be constant during all the occultation. It is provided in this order: FP1, FP2.

Offset between thermistor and CCD arrays temperature

Offsets directly read from the Calibration auxiliary product. These offsets may be used to estimate the CCD temperature. They are currently not used in the Level 1b processing which uses directly the thermistor temperature.

Sun coordinates in the geocentric equatorial inertial system

The Sun coordinates are provided for the middle of the first measurement only.

Nominal wavelength assignment

Nominal wavelength assignment of the transmission and covariance spectra written in this product. The limb spectra can be directly associated with this spectral assignment because they are not affected by pointing errors. The spectral assignment of the transmission and covariance must take the spectral shift parameter into account. This quantity, provided for each measurement, must be added to the nominal wavelength assignment to obtain the effective wavelength assignment of each spectrum. Note that if the spectrum resampling process has been performed on the nominal spectral grid, then the effective wavelength assignment is identical to the nominal assignment (in this case, the spectral shift is equal to 0 for all the measurements in the product). The nominal wavelength assignment is a global vector.

The nominal wavelength assignment of the CCD columns of each spectrometer CCD array is coded in the Level 1b product on 32 bits with an accuracy of 1.e-6 nm. The spectra provided in the Level 1b product are always in increasing spectral order CCD per CCD whatever the original spectral orientation of the CCD is (for example, the samples of the VIS CCD (SPA2) are re-ordered). There may be configurations where there is a spectral overlap between the UV and the VIS CCDs. In this case, the global vector is not in increasing order.

Effective number of star spectra used for the computation of the reference star spectrum

This is the number of star spectra that have been averaged to obtain the reference star spectrum. A maximum number of spectra is allowed as well as a minimum altitude of the tangent point to be sure that the observation is actually outside the atmosphere.

Reference star spectrum

The reference star spectrum is computed by averaging several star spectra outside the atmosphere at the beginning of the occultation. The current nominal value is 10 spectra. Note that the 2 first spectra (see the PCD at measurement level) are not used for pointing instability reason.

It is given in electrons and must be converted into physical units (ph/s/cm2/nm) by multiplying the electron values by the conversion factor inferred from the radiometric sensitivity curve (star) provided as a LUT.

Reference atmospheric density profile

These variables fully describe the reference density profile used during the ray tracing computation.

  •  Size of the reference atmospheric profile
  •  First altitude of the profile
  •  Altitude discretisation
  •  Reference atmospheric density profile

The reference atmospheric density profile is computed in the Atmosphere preparation process.

 

Definitions related to DS

Full transmission spectra and associated covariance

The transmission spectra are computed as the ratio of each measured star spectrum by the reference star spectrum, computed from star spectra measured outside the atmosphere. It is said full because it is the actually measured transmission, not corrected, neither from refraction effects (dilution, scintillation, chromatic refraction) nor from the dynamic PSF. The covariance (here, in fact, the variance of each pixel signal) is computed from analysis of S/N ratio. There is one transmission spectrum and one covariance spectrum per spectrometer measurement. The transmission and covariance spectra are global vectors (see the "Number of points of the spectra" field).

Scaled estimated central background

Estimation of the background signal measured by the target band. The estimated central background spectra values are coded on 16 bits. Due to the high variations of these spectra with the altitude, the coding is dynamic and uses an offset and a gain for each measurement. The offset and the gain are stored in the Auxiliary data ADS of the product. For each measurement, the offset is computed as the minimum value of the spectrum to store (fmin) and the gain is computed as 65535/(fmax-fmin) where fmin (resp. fmax) is the minimum (resp. the maximum) value of the spectrum for the current measurement.

Note that the background spectra are divided by the PRNU correction factor before being scaled.

The following table presents the accuracy (in number of electrons) for several ranges of background spectra:

 

Range (in e)

Accuracy (LSB)

100

< 0.1 e-

1000

< 0.1 e-

10000

0.2 e-

30000

0.5 e-

100000

1.5 e-

300000

4.6 e-

500000

7.6 e-

1000000

15.0 e-

 

To decode the value read in the product, one must compute: Background (in electrons) = Offset + Background Code / Gain

The decoded spectra values obtained in electrons must then be converted into physical units (ph/s/cm2/nm/sr) by multiplying them by the conversion factor inferred from the radiometric sensitivity curve (star) provided as a LUT.

Error bar for the estimated central background

The error bars for the estimated central background data are computed as the shot noise of the background signal and expressed in relative percentage. The maximum error bar is set to 6500%.

for each measurement f = 1, fsize
for each CCD column: k = 1, ksize
   if (  #0 ) then
  
   else
  
   endif
   /* truncate the error to 6500 percent (coding limit) */
   if (  > 6500. )    = 6500.
   end for k
end for f

Where fsize is the number of spectrometer measurements and ksize is the total number of transmitted pixels (global vectors)

They are stored in the product in 16 bits. They represent the relative percentage of error for each estimated central value. This coding allows an accuracy of at least 0.1 percent.

Photometers engineering data

Fast photometer samples converted into electrons and corrected from dark charge, straylight, background contribution and flat-field.

Error bar for the photometers engineering data

The error bar vector (50 elements) is made of 25 couples of (mean value, standard deviation) of the photometer shot noise. For each spectrometer measurement, 500 photometer samples are available. They are split into 25 sets. For each set, the mean value of the measurement shot noise and its standard deviation (based on 20 samples) are computed and are written in the product.

PCD at sample level (SP)

The meaning of the possible PCD values for the spectrometers stored in this MDS is detailed in section 5.1.3.2.2

PCD at sample level (FP)

The meaning of the possible PCD values for the photometers stored in this MDS is detailed in section 5.1.3.2.2

SATU mispointing angles

SATU data read from the Level 0 products and converted into angles (X and Y directions).

SFA angles measurements

SFA data read from the Level 0 products and converted into angles (azimuth and elevation).

Spectral shift of the star spectra

This quantity, provided for each measurement, must be added to the nominal wavelength assignment to obtain the effective wavelength assignment of each spectrum. Note that if the spectrum resampling process has been performed on the nominal spectral grid, then the effective wavelength assignment is identical to the nominal assignment (in this situation, the spectral shift is equal to 0 for all the measurements of the product).

Offset and gain for the background spectra coding

For each measurement, the offset is computed as the minimum value of the background spectrum to store (fmin) and the gain is computed as 65535/(fmax-fmin) where fmin (resp. fmax) is the minimum (resp. the maximum) value of the background spectrum for the current measurement.

This offset and this gain must be used to decode the values of the background spectra written in the product for both before and after straylight and vignetting corrections. To decode the value read in the product, one must compute:

Background (in electrons) = Offset + Background Code / Gain

There is one couple of gain and offset values per spectrometer measurement.

PCD at measurement level

The meaning of the possible flag values stored in this ADS is detailed in section 5.1.3.2.4.

Latitude, longitude and altitude of the spacecraft

Spacecraft location computed during the geolocation processing. These quantities are provided at the beginning and during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing). The altitude is expressed in meters.

Latitude, longitude and altitude of the tangent point and associated errors

Tangent point location and associated errors computed during the geolocation processing. They are given for the reference wavelength for ray tracing computation (equal to 500 nm and given in the Occultation data GADS). These quantities are provided at the beginning and during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing). The altitude is expressed in meters.

Distance spacecraft - tangent point

Distance between the satellite and the tangent point computed during the geolocation processing. It is given for the reference wavelength for ray tracing computation (equal to 500 nm and given in the Occultation data GADS). This quantity is provided at the beginning and during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).The distance is expressed in meters.

Instrument pointing direction

Instrument pointing direction (azimuth and elevation) in the GOMOS frame. These values are computed in the geolocation process and do not use the SFA angles but only the geometrical characteristics of the measurements: satellite and star location, GOMOS frame definition, orbit properties.

Virtual star direction in the quasi-true of date frame

Virtual star direction in the quasi-true of date frame provided at the beginning and during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing ).

Number of nodes of the ray tracing

The ray path is computed during the ray tracing / geolocation process. The ray is made from 3 nodes outside the atmosphere up to 150 nodes inside. If more than 150 nodes have been necessary during the processing, the ray path is resampled in order to reduce the number of nodes. This field indicates how many values are significant in the following arrays.

Index of the tangent point in the list of the ray tracing node

Index of the tangent point in the list of the ray tracing node. There is one index per spectrometer measurement (i.e. per ray tracing). This index must be used to locate the tangent point information in the ray tracing output arrays.

Interpolation factors P and Q for the law delta (lambda)

These factors may be used to compute the chromatic deviation for a specified wavelength as: deviation expressed in radians is equal to Q+P*EDLEN(l) where EDLEN is the Edlen's function, with λ expressed in nm:

where the factor 1.00062 stands for temperature and pressure correction.

These parameters are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Interpolation factors P and Q for the law h0 (lambda)

These factors may be used to compute the tangent point altitude at any wavelength in the spectral range of the spectrometers. The altitude expressed in meters is equal to Q+P*EDLEN(l) where EDLEN is the Edlen's function described above.

Latitude, longitude and altitude of the ray tracing grid nodes

Location of the ray path nodes computed during the geolocation process. The tangent point is always included in these nodes. These parameters are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Air density and atmospheric pressure at the tangent point

These values are given only for the tangent point. They are computed from the ECMWF data and MSIS90 model. The air density is expressed in cm-3 and the atmospheric pressure is expressed in Pa. They are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Air temperature at the ray tracing grid nodes

Air temperature at the ray path nodes computed by using the data from the ECMWF data and MSIS90 model. The tangent point is always included in these nodes. These values are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Sun-zenith angle at the spacecraft location

Angle computed at the middle of each spectrometer acquisition

Sun-zenith angle at the tangent point location

Angle computed at the middle of each spectrometer acquisition

Sun-azimuth angle at the tangent point location

Angle computed at the middle of each spectrometer acquisition

Apparent altitude of the central background

Altitude computed at the middle of each spectrometer acquisition

 

5.1.3.2.2 PCD and flags: PCD at Sample Level (SP)

The flags at sample level set during the Level 1b processing are combined together in a 16-bits integer before being written in the Level 1b transmission product (corresponding field in the transmission MDS) and in the limb product (corresponding field in the MDS). The way this combination is done is described in the following table.

 

Bit #

Flag name

Comments

0

FLsat(k,L,f)

saturation flag for the lower band (0: no saturation, 1: saturation)

1

FLsat(k,C,f)

saturation flag for the central band (0: no saturation, 1: saturation)

2

FLsat(k,U,f)

saturation flag for the upper band (0: no saturation, 1: saturation)

3

FLbad(k,L)

bad pixel flag for the lower band (0: no bad pixel in the band, 1: at least one bad pixel in the band)

4

FLbad(k,C)

bad pixel flag for the central band (0: no bad pixel in the band, 1: at least one bad pixel in the band)

5

FLbad(k,U)

bad pixel flag for the upper band (0: no bad pixel in the band, 1: at least one bad pixel in the band)

6

FLcr(k,L,f)

cosmic ray flag for the lower band (0: no cosmic ray detected in the data, 1: cosmic ray has been detected in the data)

7

FLcr(k,C,f)

cosmic ray flag for the central band (0: no cosmic ray detected in the data, 1: cosmic ray has been detected in the data)

8

FLcr(k,U,f)

cosmic ray flag for the upper band (0: no cosmic ray detected in the data, 1: cosmic ray has been detected in the data)

9-10

FLbg(k,f)

from the spectrometers central background estimation processing
0: central background is computed with no flagged samples
1: less than 25 percent of flagged samples
2: less than 50 percent of flagged samples
3: more than 50 percent of flagged samples

11-12

FLft(k,f)

from the transmission and covariance computations processing
0: no problem identified
1: reference star spectrum value is zero
2: one of the spatial band is saturated

13

FLoff(k)

from the wavelength assignment processing
0: pixel is in a valid spectral range
1: pixel is in an invalid spectral range

14

FLrsp(k,f)

from the reference star spectrum and transmission computation processings
0: no problem
1: current data computed with flagged data

15

 

Not used

Note: bit 0 is the lowest significant bit of the byte (bits are counted from right).

5.1.3.2.3 PCD and Flags:PCD at Sample Level (FP)

As for the spectrometer flags at sample level, the photometer flags at sample level are combined together in a 16-bits integer before being written in the Level 1b product (corresponding field in the transmission MDS). The way this combination is done is described in the following table.

Bit #

Flag name

Comments

0

photometers saturation flags (0: no saturation, 1: saturation)

1-15

 

Not used

Note: bit 0 is the lowest significant bit of the byte (bits are counted from right).

5.1.3.2.4 PCD and Flags:PCD at measurement level

Each flag is affected to a whole spectrometer measurement i.e. gives global information concerning all the spatial bands, all the CCD columns of the spectrometers and all the photometer samples during the measurement. There are two kinds of flags at measurement level: the first set is made of flags explicitly set at measurement level such as the data valid flag (read in the GOMOS packets), the flags indicating if SATU outputs have been used for correction or if the star spectrum was outside the central band ... The second set of flags is built from the flags at sample level and can be seen as a flag summary at measurement level.

The PCD at measurement level are written in the auxiliary data ADS of the Level 1b transmission product and in the ADS of the Level 1b limb product.

 

Flags explicitly set at measurement level:

 

Variable

Description

Range - References

FLAGDV(f)

data valid flag (measurement level)

from the SDP extraction processing
0: anomaly
1: time-out
3: fully successful
9: missing packet

FLAGtime(f)

datation flag

from the datation processing
0: no datation problem
1: problem in the datation processing
2: invalid measurement (FLAGDV(f)=0 or 1)
9: missing packet

FLAGrt(f)

ray-tracing flag

from the geolocation processing
0: no problem
1: wrong pointing (not in the direction of the atmosphere)
2: ray path across the Earth

FLAGgl(f)

geolocation flag

from the geolocation processing
0: no problem
1: geolocation problem

FLAGirv(f)

vignetting correction flag

from the spectrometers vignetting correction processing
0: no vignetting
1: vignetting occurs for this measurement

FLAGout(f)

indicates if the star spectrum falls outside the central band during the measurement

from the star spectrum computation
0: star spectrum is inside the central band
1: centre line of star spectrum image is outside the central band

 

indicates the number of fast photometers saturation during the spectrometer measurement

from the photometers saturated samples processing
fc: FP1, FP2

FLAGstab(f)

indicates instable measurements

The number of instable measurements at the beginning of the occultation (fstab) is read from the Calibration auxiliary product.

For all measurements before fstab (including it), FLAGstab(f) is set to 1 and for all following measurements FLAGstab(f) is set to 0.

FLAGmod(f)

modulation correction flag

from the spectrometer demodulation processing

0: no problem

+1: problem of structure for SPA1 upper band

+2: problem of structure for SPA1 lower band

+4. inconsistent upper and lower band structures for SPA1

+10: problem of structure for SPA2 upper band

+20: problem of structure for SPA2 lower band

+40. inconsistent upper and lower band structures for SPA2

Note that the flag values between SPA1 and SPA2 may be mixed, ie flag values like 11, 21, 32, 23, 41, 24 can exist

FLAGUC(f)

ratio between the averaged upper band signal and the averaged star signal over the SPA2 spectral range

from the star spectrum computation
n<65535: computed ratio expressed in %
65535: star signal is equal to 0

 

Note: FLAGDV(f) and FLAGDB are read in the GOMOS packets. They are not generated by the Level 1b processing.

 

Flags at measurement level computed from flags at sample level:

The following flags at measurement level are computed as the sum over the number of CCD columns of the corresponding flags at sample level.

 

Variable

Description

Range - References

FLAGsat(f)

number of saturated samples per measurement

from the spectrometers saturated samples processing

FLAGcr(f)

number of cosmic ray detections per measurement

from the spectrometers cosmic rays detection processing

FLAGbg(f)

number of background flagged data per measurement

from the estimated central background processing

FLAGft(f)

number of samples per measurement for which

FLft(k,f) has been raised

from the transmission and covariance computations processing

FLAGsum(f)

PCD at measurement level (summary)

16 integer values - filled with the various flags at each measurement (see next table); written in the Level 1b and Limb products

 

For each spectrometer measurement, each component of the working array FLAGsum(f) (16 integer values) is set to one of the flags at measurement level as shown in the following table. The array FLAGsum(f) is written in each Auxiliary data ADSR of the Level 1b and Limb products.

 

The order of the flags is described in the table below.

 

Array index

Flag name

Comments

FLAGsum(f)(1)

FLAGDV(f)

data valid flag

FLAGsum(f)(2)

 

not used

FLAGsum(f)(3)

FLAGtime(f)

datation flag

FLAGsum(f)(4)

FLAGrt(f)

ray-tracing flag

FLAGsum(f)(5)

FLAGgl(f)

geolocation flag

FLAGsum(f)(6)

FLAGsat(f)

saturation flag (SP)

FLAGsum(f)(7)

FLAGcr(f)

cosmic ray flag

FLAGsum(f)(8)

FLAGirv(f)

vignetting flag

FLAGsum(f)(9)

FLAGbg(f)

background flag

FLAGsum(f)(10)

FLAGout(f)

star spectrum out of band

FLAGsum(f)(11)

FLAGft(f)

transmission flag

FLAGsum(f)(12)

saturation flag (FP1)

FLAGsum(f)(13)

saturation flag (FP2)

FLAGsum(f)(14)

fstab

stability flag

FLAGsum(f)(15)

not used

set to 0

FLAGsum(f)(16)

not used

set to 0

 

5.1.3.2.5 PCD and Flags: PCD at occultation level

Each flag characterises the whole occultation. There are three kinds of flags at occultation level:

  • the first set is made of flags explicitly set at occultation level (e.g. dark/bright limb flag, star reference spectrum fatal flag...).
  • the second set of flags is built from the flags at measurement level (e.g. saturated samples, cosmic rays...) and can be seen as a flag summary.
  • the third set of flags is built from flags at sample level (e.g. samples containing bad pixels).

 

Flags explicitly set at occultation level:

 

Variable

Description

Range - References

PCDist

PCD for internal straylight correction processing

from the internal straylight correction processing

0: the internal straylight correction has been applied

1: the internal straylight correction has not been applied

PCDearth

PCD for external earth straylight correction processing

from the external straylight correction processing

0: the external earth straylight correction has been applied

1: the external earth straylight correction has not been applied

PCDsun

PCD for external sun straylight correction processing

from the external straylight correction processing

0: the external sun straylight correction has been applied

1: the external sun straylight correction has not been applied

PCDslit

PCD for slit transmission correction processing

from the slit transmission correction processing

0: the slit transmission correction has been applied

1: the slit transmission correction has not been applied

PCDref

PCD indicating that the reference star spectrum is computed with a small number of star spectra

from the reference star spectrum computation processing

0: no problem

1: the reference star spectrum has been computed with a small number of measurements

2: no valid measurements in the occultation to compute the reference star spectrum (see PCDsbd and PCDfatal)

PCDsdb

PCD indicating if the star spectrum has been read in the Stellar Spectra Database

from the reference star spectrum computation processing

0: the reference star spectrum is computed from the measurements

1: the reference star spectrum is read in the Stellar spectra database

2: the reference star spectrum has not been found in the Stellar spectra database (see PCDfatal)

PCDfatal

PCD indicating that the reference star spectrum has not been computed

from the reference star spectrum computation processing

0: the reference star spectrum has been computed

1: the reference star spectrum has not been computed (no transmission computed)

PCDSATU

PCD indicating that the SATU data are used for flat-field correction

from the Level 1b processing configuration database (same value as OKSATU)

0: the SATU data are not used

1: the SATU data are used

PCDlv0

PCD read from the Level 0 product

from the SDP extraction processing

0: standard occultation

1: first part of a tangent occultation

2: last part of a tangent occultation

PCDatm

PCD for atmosphere file

from the preparation of the Atmosphere model

54: one ECMWF file used. Time record is included in the occultation period.

102: one ECMWF file used. Time record before the beginning of the occultation ( Dt Δ≤ 24h.

103: one ECMWF file used. Time record after the beginning of the occultation (Dt24h).

106: one ECMWF file used. Only one time record in the validity interval (Dt24h).

155: two ECMWF files used.

201: no ECMWF file available. Use MSIS model alone.

202: only old ECMWF files (Dt>24h). Use MSIS model alone.

203: only future ECMWF files (Dt>24h). Use MSIS model alone.

206: no ECMWF file available in the validity interval (Dt>24h). Use MSIS model alone.

In case of no convergence of the ray tracing, the PCDatm flag is increased by 10. Example, an initial value of 106 of PCDatm will be actually set to 116 in case of no convergence.

PCDdc

PCD for dark charge correction

from the dark charge correction processing

indicates the dark charge correction processing applied and if problems have occurred

0: dark charge map used

1: first measurements used

2: no dark charge correction applied

11: first measurements not available - DC correction with DC maps

12: first measurements not available and missing 1st packet - DC correction with DC maps without temperature dependence

21: dark charge map computed from DSA observation has been used

okback

Background correction activation switch

from the level 1b processing configuration or modified by the background correction processing.

indicates which background correction has been effectively performed.

Note that in full dark limb condition, no correction is applied to SPA samples, whatever the value of the activation switch read from the level 1b processing configuration.

0: no correction

1: background correction applied (linear interpolation)

2: background correction applied (exponential interpolation)

3: background correction applied (general method)

 

Flags at occultation level for the photometers:

 

Variable

Description

Range - References

PCD for dark charge correction of the photometers samples

from the photometers dark charge correction

indicates the dark charge correction processing applied (0:DC correction performed; 1:no DC correction)

PCD for saturation: number of photometers saturated samples

from the photometers bad pixels processing

 

Flags at occultation level computed from flags at measurement level:

The following flags at occultation level are computed as the sum over the number of measurements of the corresponding measurement flags.

 

Variable

Description

Range - References

PCDfvalid

No valid data flag

set to 1 if PCDDV = fsize

set to 0 otherwise

Nerr

Number of source packets containing errors

from the SDP extraction processing

number of invalid or missing measurement

PCDDB

PCD for dark/bright limb conditions

from the SDP extraction processing

0: dark limb
1: bright limb

PCDillum

Observation illumination condition

from the geolocation processing. This is the actual illumination of the observation, computed from geometrical configuration between the Sun, the satellite and the GOMOS instrument pointing direction.

            0: full dark limb condition

            1: bright limb condition
            2: pure twilight condition
            3: straylight condition
            4: twilight+straylight condition

This flag is not dependent on the dark/bright limb flag that is read in the Level 0 product; as this latter comes from the mission scenario and may not fully reflect the actual observation conditions.

PCDDV

PCD of the SDP extraction processing

from the SDP extraction processing

number of invalid or missing measurements

PCDtime

PCD for datation: indicates the number of measurements where flags have been raised

from the datation processing

PCDrt

PCD for ray tracing: indicates the number of measurements with invalid ray path

from the geolocation processing

PCDgl

PCD for geolocation: indicates the number of measurements performed outside the atmosphere

from the geolocation processing

note: set to 1000 if the occultation is completely performed outside the atmosphere

PCDsat

PCD for saturation: indicates the number of measurements containing saturated samples

from the spectrometers saturated samples processing

number of measurements where saturated samples have been detected

PCDcr

PCD for cosmic rays: indicates the number of measurements where cosmic rays has been detected

from the cosmic rays detection processing

number of measurements where cosmic rays have been detected

PCDmod

PCD for modulation correction processing

from the modulation correction processing
number of measurements where the modulation correction has not been applied

PCDirv

PCD for vignetting correction processing

from the vignetting processing

number of measurements where the vignetting correction has been applied

PCDbg

PCD of the central background computation 

from the estimated central background processing

number of measurements where FLAGbg(f) has been raised

PCDout

PCD for flat-field correction: indicates the number of measurements where the star falls outside the central band

from the star signal computation processing

number of measurements where part of the star spectrum is outside the central band

PCDft

PCD indicating that a problem has occurred during the computation of the transmission or of the covariance

from the transmission and covariance computations processing

number of measurements where a problem has been detected during the full transmission computation

 

Flags at occultation level computed from flags at sample level:

The following flags at occultation level are computed as the sum over the number of CCD columns of the corresponding flags at sample level.

 

Variable

Description

Range - References

PCDbad

PCD for bad pixels

from the spectrometers bad pixels processing

number of bad pixels (same value for all the measurements)

  

The PCDs at occultation level are written in the Summary quality GADS of the Level 1b transmission products, of the Level 1b limb products, of the Level 2 atmospheric constituents products, and of the Level 2 residual extinction products.

5.1.3.3 Limb product

Most of the fields included in the limb product are also included in the GOMOS level 1b products. Only the different fields are explained in the table below. For the other fields, please refer to the previous level 1b description chapter.

 

Definitions related to DS

Summary quality

The meaning of the possible flag values stored in the Summary Quality GADS is detailed in section 5.1.3.2.1

Radiometric sensitivity curve (background)

This curve has been built from the radiometric sensitivity of the instrument measured during its on-ground characterisation and updated in-flight. It allows the conversion of the background spectra expressed in electrons in the product into physical units (ph/s/cm2/nm/sr). It is given as a LUT of conversion factors. A linear interpolation of the conversion factor is needed to use this curve for any sample of the background spectra.

Scaled upper & lower background spectra without straylight correction

Estimation of the background signal measured by the upper and lower bands before applying the straylight and the IR-vignetting corrections. The estimated background spectra values are coded on 16 bits. Due to the high variations of these spectra with the altitude, the coding is dynamic and uses an offset and a gain for each measurement. The offset and the gain are stored in the Auxiliary data ADS of the product. For each measurement, the offset is computed as the minimum value of the spectrum to store (fmin) and the gain is computed as 65535/(fmax-fmin) where fmin (resp. fmax) is the minimum (resp. the maximum) value of the spectrum for the current measurement. To decode the values read in the product, one must compute: Background (in electrons) = Offset + Background Code / Gain

Note that the background signal is divided by the average PRNU map before being scaled.

 

The following table presents the accuracy (in number of electrons) for several ranges of background spectra:

 

Range (in e)

Accuracy (LSB)

100

< 0.1 e-

1000

< 0.1 e-

10000

0.2 e-

30000

0.5 e-

100000

1.5 e-

300000

4.6 e-

500000

7.6 e-

1000000

15.0 e-

 

The decoded spectra values obtained in electrons must then be converted into physical units (ph/s/cm2/nm/sr) by multiplying them by the conversion factor inferred from the radiometric sensitivity curve (star) provided as a LUT.

Scaled upper & lower background spectra after straylight and IR vignetting corrections

Estimation of the background signal measured by the upper and lower bands after applying the straylight and IR-vignetting corrections. The estimated background spectra values are coded on 16 bits. Due to the high variations of these spectra with the altitude, the coding is dynamic and uses an offset and a gain for each measurement. The offset and the gain are stored in the Auxiliary data ADS of the product. For each measurement, the offset is computed as the minimum value of the spectrum to store (fmin) and the gain is computed as 65535/(fmax-fmin) where fmin (resp. fmax) is the minimum (resp. the maximum) value of the spectrum for the current measurement. To decode the values read in the product, one must compute: Background (in electrons) = Offset + Background Code / Gain

Note that the background signal is divided by the average PRNU map before being scaled.

The following table presents the accuracy (in number of electrons) for several ranges of background spectra:

 

 Range (in e)

Accuracy (LSB)

100

< 0.1 e-

1000

< 0.1 e-

10000

0.2 e-

30000

0.5 e-

100000

1.5 e-

300000

4.6 e-

500000

7.6 e-

1000000

15.0 e-

 The decoded spectra values obtained in electrons must then be converted into physical units (ph/s/cm2/nm/sr) by multiplying them by the conversion factor inferred from the radiometric sensitivity curve (star) provided as a LUT.

Error bar for the upper & lower background spectra after straylight and IR vignetting corrections

The error bars for the upper and lower background spectra after straylight and vignetting corrections are computed as their square root and are expressed as a relative percentage. The maximum error bar for the upper and lower background spectra is set to 255%.

for each measurement f = 1, fsize
            for each CCD column: k = 1, ksize
                        if (  #0 ) then
                                  
                        else
                                  
                        endif
                        /* truncate the error to 255 percent (coding limit) */
                        if (  > 255. )   = 255.
            end for k
end for f

Where fsize is the number of spectrometer measurements and ksize is the total number of transmitted pixels (global vectors)

PCD at sample level

The meaning of the possible PCD values stored in this MDS is detailed in section 5.1.3.2.2.

Offset and gain for the background spectra coding

For each measurement, the offset is computed as the minimum value of the background spectrum to store (fmin) and the gain is computed as 65535/(fmax-fmin) where fmin (resp. fmax) is the minimum (resp. the maximum) value of the background spectrum for the current measurement.

This offset and this gain must be used to decode the value of the background spectra written in the product for both before and after straylight and vignetting corrections. To decode the value read in the product, one must compute:

Background (in electrons) = Offset + Background Code / Gain

There is one couple of gain and offset values per spectrometer measurement.

PCD at measurement level

The meaning of the possible PCD values stored in this MDS is detailed in section 5.1.3.2.4.

 

5.1.3.4 Level 2 product

 Definitions related to GADS

Summary Quality

The meaning of the possible flag values stored in the Summary Quality GADS is detailed in section 3.3.2.

Level 1b PCD check

This PCD is initialised when the Level 1b product is read, following the logic:

First of all, read the summary quality GADS in the Level 1b product:

C1 : no valid data
test if PCDfvalid=1 (field 1 of the GADS)

C2 : second part of a tangent occultation
test if PCDlv0=2 (field 12 of the GADS)

C3 : geolocation is not valid
test if PCDgl=1000 (field 20 of the GADS)

C4 : reference star spectra missing, i.e. transmission terms not valid for the whole occultation
test if PCDfatal>0 (field 8 of the GADS)

Initialise PCDlv1=0

If C1 is true set PCDlv1 to 1
If C2 is true set PCDlv1 to 2
If C3 is true set PCDlv1 to 3
If C4 is true set PCDlv1 to 4

Chromatic refraction mode for the measured transmission

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. The possible values are:

0 : no correction
1 : correction is performed

Chromatic refraction mode for the transmission model (second spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the auxiliary product: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: no correction
1: correction is performed

Chromatic refraction mode for the transmission model (third and further spectral inversions)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the auxiliary product: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: no correction
1: correction is performed

Instrument function mode for the transmission model (second spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the auxiliary product: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: no correction
1: correction is performed

First altitude where the ratio U/C is greater than 25%

This value is computed from the PCD at measurement level provided in the Level 1b product (see Section 2.1.1.1.3).

It is specified in km, truncated to the nearest integer value. If the ratio is lower than 25% at all altitudes, this altitude is set to 0.

Vertical inversion mode

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. The possible values are:

2: linear (only one mode is available in the processor)

Smoothing mode (after the spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Smooth or not the density column used by the vertical inversion. If the smooth mode is not activated only the flagged value of the density column are filled.

0: no smoothing
1: Gaussian filter

2: Tikhonov's regularisation

Time mode for the transmission model (second spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the database value: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: zero order term
1: second order term

Time mode for the transmission model (third and further spectral inversions)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the database value: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: zero order term
1: second order term

 

Number of iterations for the main loop (also called q loop)

Number of main loop iterations actually performed.

 

Number of iterations for the inversion loop (also called p loop)

Number of inversion loop iterations actually performed.

Number of points in profile column densities where chi2 > chi2_warn

chi2_warn is a parameter read in the Level 2 processing configuration auxiliary product (LUT versus altitude).

 

Number of flagged points in profile for the column densities

Number of points where the corresponding PCD (tangent line density PCD) is not equal to 0.

Number of flagged points in profile for the local densities

Number of points where the corresponding PCD (local density PCD) is not equal to 0.

 

 

Definitions related to DS

Local density product

This is the result of the vertical inversion of line densities, assuming local spherical symmetry. The local density of each species is expressed in cm-3. One value is provided for each measurement. The datation and the geolocation of each point of the local density profile is also provided in the product.

Standard deviation of the local density product

Errors are estimated throughout the processing chain and are propagated along the chain to the final data products. It is assumed that the error statistics follow a normal Gaussian distribution. Values given correspond to 1. The error bar associated with the local density of the different species is expressed in % of the local density. The maximum value of the error bar is set to 6553.5%.

PCD summary (for the local density profiles)

The PCD summary contains flags dedicated to the validity of the outputs of the vertical inversion, i.e. the retrieval of the local density profiles. After vertical inversion, local densities are checked for flag setting. There is one flag per acquisition and per species: parameter PCDV(i,j), where index j denotes the acquisition number, and index i denotes the species among the following species list in this order: O3, NO2, NO3, air, O2, H2O, OClO.

Tangent line density

Tangent line density of each species (expressed in cm-2). One value is provided for each measurement. The tangent line density profiles are computed during the spectral inversion process.

Standard deviation for the tangent line density

Errors are estimated throughout the processing chain and are propagated along the chain to the final data products. It is assumed that the error statistics follow a normal Gaussian distribution. Values given correspond to 1. With the current operational processor (IPF 5.00), an empirical error estimate has been added after inversion to the error bar of O3 line densities to account for the effects of turbulence (incomplete  scintillation correction). The error bar associated with the tangent line density of the different species is expressed in % of the tangent line density. The maximum value of the error bar is set to 6553.5%.

PCD summary (for the tangent line density profiles)

The PCD summary contains flags dedicated to the validity of the outputs of the spectral inversion, i.e. the retrieval of the tangent line density profiles. After spectral inversion, tangent line densities are checked for flag setting. There is one flag per acquisition and per species: parameter PCDV(i,j), where index j denotes the acquisition number, and index i denotes the species among the following species list in this order: O3, NO2, NO3, air, O2, H2O, OClO.

Aerosol product

The wavelength dependence of aerosol extinction may vary, even over one occultation. A polynomial expression is used to describe this wavelength dependence of the aerosol extinction; for a polynomial in degree n:

 

 

For each measured transmission, it is calculated:

o       the extinction coefficient at the reference wavelength , corresponding to d0, given in km-1

o       the other spectral parameters corresponding to the coefficients d1(z) to dn(z) in the polynomial expression, given in nm-1.km-1, ..., nm-n.km-1

The reference wavelength value is stored in the SPH of the product ("Reference wavelength used for the ray tracing"); it is equal to 500 nm. It is also given in the SPH of the residual extinction product and in the Occultation data GADS of the Level 1b product.

 

The extinction coefficient at  is given in the dedicated MDSR "Extinction coefficient at " and as the first spectral parameter of the MDS "Spectral parameters of the extinction coefficients". In the case of a polynomial in degree 2 as in the current operational processor (IPF 5.00), two other coefficients are given in the MDS "Spectral parameters of the extinction coefficients": d1 (nm-1.km-1) and d2 (nm-2.km-1).

 

The wavelength dependence of the aerosol optical thickness may be described as:

 

 

The MDS "Tangent integration extinction profile at λ=λref " stores the value τ calculated at λ=λref (dimensionless). The MDS "Spectral parameters of tangent integrated extinction profile" stores the coefficients r0 to rn, retrieved by the spectral inversion processing; r0 is the integrated retrieved density column of aerosols given in cm-2; the other parameters r1 to rn are given in cm-2.nm-1, ... cm-2.nm-n. In the case of a polynomial in degree 2 as in the current operational processor (IPF 5.00), three coefficients are stored: r0 (dimensionless), r1 (cm-2.nm-1) and r2 (cm-2.nm-2).

Extinction coefficient

The extinction coefficient is given at the reference wavelength, in km-1. The reference wavelength is given in the SPH of the product and in the Occultation data GADS of the Level 1b product; it is equal to 500 nm.

Standard deviation of the extinction coefficient

The values of the standard deviation are expressed in % and correspond to 1. An empirical error estimate has been added after inversion to account for the effects of turbulence (incomplete scintillation correction).

Spectral parameters of the extinction coefficients

Coefficients of the polynomial expression used to describe the wavelength dependence of the aerosol extinction. The first spectral parameter is the extinction coefficient at the reference wavelength, given in km-1; for a polynomial in second degree, two other parameters are given, in nm-1.km-1 and in nm-2.km-1 respectively.

Standard deviation of the spectral parameters of the extinction coefficients

The values of the standard deviation are expressed in % and correspond to 1.

 

Tangent integrated extinction profile

The tangent integrated extinction (aerosol optical thickness, dimensionless) profile is given at the reference wavelength. The reference wavelength is given in the SPH of the product and in the Occultation data GADS of the Level 1b product; it is equal to 500 nm.

Standard deviation of the tangent integrated extinction profile

The values of the standard deviation are expressed in % and correspond to 1.

 

Spectral parameters of the tangent integrated extinction profile

Coefficients of the polynomial expression used to describe the wavelength dependence of the aerosol optical thickness. The first spectral parameter is the integrated retrieved density column of aerosols given in cm-2; for a polynomial in second degree, two other parameters are given, in cm-2.nm-1 and cm-2.nm-2 respectively.

Standard deviation of the spectral parameters of the tangent integrated extinction profile

The values of the standard deviation of the polynomial coefficients of the tangent integrated extinction profile are expressed in % and correspond to 1.

PCD summary

The PCD summary provides the spectral and vertical PCD (first and seventh values) of the extinction coefficient at the reference wavelength; the other values are set to 0.

Turbulence product

From the analysis of fast photometers, the tangent altitude of LOS, and the temperature and density vertical profiles at High Resolution (40 Hz) are retrieved. They are basically derived from the time delay between the peaks of red and blue Fast Photometers, when the time delay between the two photometer signal is significant enough to be analysed. There is no high resolution profiles when the occultation is made in bright limb.

Tangent altitude including fluctuations

Altitude of the high resolution temperature profile.

High resolution temperature vertical profile

High resolution T vertical profile obtained by analysis of the photometer signals. The output frequency of the HRTP is 40Hz. This corresponds to 20 values for each processed measurement during the Level 2 processing (frequency 2Hz). There is no high resolution temperature profile when the occultation is made in bright limb or if the time delay between the two photometer signals is not significant enough to be analysed.

Error bar of the high resolution temperature profile

The values of the error bar are expressed in % and correspond to 1. The maximum error bar is set to 6500%.

High resolution vertical profile of the local density

High resolution vertical profile of air local density obtained by analysis of the photometer signals. The output frequency of the HRTP is 40Hz. This corresponds to 20 values for each processed measurement during the Level 2 processing (frequency 2Hz). There is no high resolution air density profile when the occultation is made in bright limb or if the time delay between the two photometer signals is not significant enough to be analysed.

Error bar of the high resolution density profile

The values of the error bar are expressed in % and correspond to 1. The maximum error bar is set to 6500%.

Geolocation and atmospheric profile

Geolocation data are the same as for Level 1b and are also provided for the reference wavelength (equal to 500 nm).

Latitude, longitude and altitude of the spacecraft

Spacecraft location read from the Level 1b product. These quantities are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).The altitude is expressed in meters.

Latitude, longitude and altitude of the tangent point and associated errors

Tangent point location and associated errors read from the Level 1b product. These quantities are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing). The altitude is expressed in meters.

Tangent point atmospheric pressure (from external model)

Pressure profile obtained by the combination of the ECMWF data in the lower part of the profile and of the MSIS90 data in the upper part of the profile (smooth transition altitude range around the pressure level 1hPa). The values of the profile are read in the Level 1b product. The values are given in Pa. They are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Note: if the ECMWF file is not available for the Level 1b processing, only the MSIS90 model is used. In this case, a specific flag is raised during the Level 1b processing and copied in the Level 2 product.

Tangent point temperature (from external model)

Temperature profile obtained by the combination of the ECMWF data in the lower part of the profile and of the MSIS90 data in the upper part of the profile (smooth transition altitude range around the pressure level 1hPa). The values are given in K. They are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Note: if the ECMWF file is not available for the Level 1b processing, only the MSIS90 model is used. In this case, a specific flag is raised during the Level 1b processing and copied in the Level 2 product.

Tangent point density (from external model)

Air density obtained by the combination of the ECMWF data in the lower part of the profile and of the MSIS90 data in the upper part of the profile (smooth transition altitude range around the pressure level 1hPa). The values are given in cm-3. They are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Note: if the ECMWF file is not available for the Level 1b processing, only the MSIS90 model is used. In this case, a specific flag is raised during the Level 1b processing and copied in the Level 2 product.

Local air density from GOMOS atmospheric profile

Local air density profile computed from the measurements. In the current operational IPF version (IPF 5.00), the vertical inversion on air is not activated, and the local density values, the standard deviation and the vertical resolution for this species are set to 0 in the products (as well as the terms related to air in the covariance matrix for local densities after the vertical inversion).

Standard deviation for the local air density

The values of the error bar are expressed in % and correspond to 1. The maximum error bar is set to 6553.5%. In the current operational IPF version (IPF5.00), all values are set to the maximum error bar.

Local temperature

Local temperature profile computed from the measurements. In the current operational IPF version (IPF 5.00), the profile values are all set to 0.

Standard deviation for the local temperature

The values of the error bar are expressed in % and correspond to 1. The maximum error bar is set to 6553.5%. In the current operational IPF version (IPF5.00), all values are set to the maximum error bar.

Accuracy estimate product

For the spectral inversion, the c2 (final best fit) is given. The 12 x 12 covariance matrix (symmetrical) is given, corresponding to the 6 gas species, 5 aerosol parameters and one spare (influence of one gas on the others).

For the vertical inversion the covariance matrix is also given (influence of one altitude on the other).

A scale factor is applied on the covariance matrix terms before recording.

Chi-2 final value

c2 final value from the spectrometer A spectral inversion.

Scale factor for the elements of the covariance matrix (spectral/vertical inversion)

The scale factor is the power of 10 to be applied to the associated covariance matrix elements in order to store/read them in the product. More details are given in the text below.

Covariance matrix for line densities after spectral inversion

The covariance matrix for the spectral inversion provides information on the influence of each gas on the others. It is a 12 x 12 matrix whose terms are in the following order: O3, NO2, NO3, air, OClO, aerosols (a maximum of 6 parameters for aerosol spectral dependency) and a spare gas. The matrix is symmetrical: only half of the matrix (78 values) is written in the product. More details on the storage of these values are given in the text below.

Covariance matrix for local densities after vertical inversion

The covariance matrix for the vertical inversion provides information on the influence of each altitude on the other. It is symmetrical. 12 species are given in the following order: O3, NO2, NO3, air, O2, H2O, OClO, aerosols and 4 spare gases. Only the diagonal terms and 6 off-diagonal terms are given in the product, as a 12 x 7 matrix for each acquisition. More details on the storage of these values are given in the text below.

 

The scale factor for the elements of the covariance matrix DS gives the power of 10 to be applied to the covariance matrix elements read in the product in order to interpret them (same value for all species):

V = Vp x 10fact

where:

V: the covariance matrix element calculated in the Level2 processing

Vp: the scaled covariance matrix element read in the product

fact: the scale factor read in the product.

 

The covariance matrix for spectral inversion is a 12 x 12 matrix whose terms [Ax,y] are in the following order: O3, NO2, NO3, air, OClO, aerosols (a maximum of 6 parameters), and a spare gas. This matrix is symmetrical, only the relevant part (78 terms) is written in the product; that is:

 

 

For x=1,12

write A(x,y)=CSIi,i',j  y ≥ x, with the species i and i' being those provided by the order x and y defined above

End For

 

The covariance matrix for vertical inversion is, by construction, symmetrical. Only the diagonal + 6 off diagonal terms are written in the product. 12 species are considered in the following order: O3, NO2, NO3, air, O2, H2O, OClO, aerosols and 4 spare gases. Thus, a matrix 12 x 7 (noted B hereafter) is written in the product for each acquisition.

 

The procedure for the creation of this part of the Level 2 product is the following:

  • One row is used for each species
  • For each species (row), the diagonal element of the original covariance matrix is written in the last column.
  • elements taken from the left of the diagonal in the original matrix are written in the first six columns
  • An exception is made for the first 6 acquisitions, where less than 6 elements are available on the left.

For these first 6 acquisitions, the elements are taken instead from the right of the diagonal in the original matrix and are inserted into the matrix B in the first six columns, but in reverse order (so that the elements closest to the diagonal in the original are also closest in the matrix B).

 

This procedure is more explicitly described below:

 

For the current acquisition j

if (j>6) then

   for i = 1,12

      B (i,7+p) = CVli,j,j+p with p = -6 to 0

   end for

else

   for i = 1,12

      B (i,7-p) = CVli,j,j+p with p = 6 to 0

   end for

end if

Note 1 : for all spare gases (or non retrieved species), B(i,7-p)=0 with p=6 to 0

Note 2 : the scaling is applied to the matrix term prior to this recording

 

Explicit illustration is given on the following pictures:

 


 

5.1.3.5 Residual extinction

 

Definitions related to the SPH

Number of processed measurements during the Level 2 processing

This is the actual number of measurements included in the product. This number is always lower than the total number of measurements of the occultation (read in the Level 1b product). It may be smaller if problem occurred (e.g. invalid data at the beginning of the occultation).

Reference wavelength used for the ray tracing

All the geolocation information is provided for this reference wavelength. For example the localisation of the ray nodes and of the tangent point node are provided for this wavelength although at one given time inside the atmosphere, and due to chromatic refraction effects, each wavelength is looking at a different altitude. It is also the reference wavelength for the aerosol products. It is equal to 500 nm.

Time shift for the ray tracing / geolocation

This time indicates the time shift between what is called 'beginning of measurement' and 'half measurement' in the ray tracing processing. Geolocation results are provided at these two times for each measurement (actual value read from the Level 1b product)

Mean wavelength of the photometer used for the scintillation correction

Value computed from the photometer spectral range and index of the photometer used for the scintillation correction.

 

 

 

Note: most of the SPH fields are identical to those of the Level 1b product and are not detailed here.

 

 

Definitions related to GADS

Summary quality

The meaning of the possible flag values stored in the Summary Quality GADS is detailed in section 3.3.2.

Level 1b PCD check

This PCD is initialised when the Level 1b product is read, following the logic:

First of all, read the summary quality GADS in the Level 1b product:

C1 : no valid data
test if PCDfvalid=1 (field 1 of the GADS)

C2 : second part of a tangent occultation
test if PCDlv0=2 (field 12 of the GADS)

C3 : geolocation is not valid
test if PCDgl=1000 (field 20 of the GADS)

C4 : reference star spectra missing, i.e. transmission terms not valid for the whole occultation
test if PCDfatal>0 (field 8 of the GADS)

Initialise PCDlv1=0

If C1 is true set PCDlv1 to 1
If C2 is true set PCDlv1 to 2
If C3 is true set PCDlv1 to 3
If C4 is true set PCDlv1 to 4

Chromatic refraction mode for the measured transmission

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. The possible values are:

0 : no correction
1 : correction is performed

Chromatic refraction mode for the transmission model (second spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the auxiliary product: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: no correction
1: correction is performed

Chromatic refraction mode for the transmission model (third and further spectral inversions)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the auxiliary product: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: no correction
1: correction is performed

Instrument function mode for the transmission model (second spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the auxiliary product: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: no correction
1: correction is performed

First altitude where the ratio U/C is greater than 25%

This value is computed from the PCD at measurement level provided in the Level1b product (see Section 2.2.3.2).

It is specified in km, truncated to the nearest integer value. If the ratio is lower than 25% at all altitudes, this altitude is set to 0.

Vertical inversion mode

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. The possible value is:

2: linear (only one mode is available in the processor)

Smoothing mode (after the spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Smooth or not the density column used by the vertical inversion. If the smooth mode is not activated only the flagged value of the density column are filled.

0: no smoothing
1: Gaussian filter

2: Tikhonov's regularisation

Time mode for the transmission model (second spectral inversion)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the database value: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: zero order term
1: second order term

Time mode for the transmission model (third and further spectral inversions)

Switch read from the Level 2 processing configuration auxiliary product and copied in the Level 2 product. Two numbers are read in the database value: first value is for the second spectral inversion and second value is for the third and subsequent spectral inversions.

0: zero order term
1: second order term

Number of iterations for the main loop (also called q loop)

Number of main loop iterations actually performed.

 

Number of iterations for the inversion loop (also called p loop)

Number of inversion loop iterations actually performed.

Number of points in profile column densities where chi2 > chi2_warn

chi2_warn is a parameter read in the Level 2 processing configuration auxiliary product (LUT versus altitude).

 

Number of flagged points in profile for the column densities

Number of points where the corresponding PCD (tangent line density PCD) is not equal to 0.

Number of flagged points in profile for the local densities

Number of points where the corresponding PCD (local density PCD) is not equal to 0.

Nominal wavelength assignment

Nominal wavelength assignment read from the Level 1b product and re-ordered in increasing wavelength value. Note that the effective spectral assignment of the transmission spectra written in the product is the combination of the nominal wavelength assignment and of the spectral grid correction.

 

 

Definitions related to DS

Transmission corrected for scintillation and dilution effects

Transmission read in the Level 1b product and corrected from the scintillation and dilution effects.

Transmission model function

The transmission model function is a forward model of the transmission with the "best fit" values of the parameters

Latitude, longitude and altitude of the spacecraft

Spacecraft location read from the Level 1b product. These quantities are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing). The altitude is expressed in meters.

Latitude, longitude and altitude of the tangent point and associated errors

Tangent point location and associated errors read from the Level 1b product. These quantities are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing). The altitude is expressed in meters.

Tangent point atmospheric pressure (from external model)

Pressure profile obtained by the combination of the ECMWF data in the lower part of the profile and of the MSIS90 data in the upper part of the profile (smooth transition altitude range around the pressure level 1hPa). The values of the profile are read in the Level 1b product. The values are given in Pa. They are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Note: if the ECMWF file is not available for the Level 1b processing, only the MSIS90 model is used. In this case, a specific flag is raised during the Level 1b processing and copied in the Level 2 product.

Tangent point temperature (from external model)

Temperature profile obtained by the combination of the ECMWF data in the lower part of the profile and of the MSIS90 data in the upper part of the profile (smooth transition altitude range around the pressure level 1hPa). The values are given in K. They are provided during the measurement, generally at half-measurement (the actual temporal shift is given by the time shift for the ray tracing).

Note: if the ECMWF file is not available for the Level 1b processing, only the MSIS90 model is used. In this case, a specific flag is raised during the Level 1b processing and copied in the Level 2 product.

Local air density from GOMOS atmospheric profile

Local air density profile computed from the measurements. In the current operational IPF version (IPF5.00), the vertical inversion on air is not activated, and the local density values are set to 0 in the products.

Spectral grid correction for the transmission model function

This spectral grid is the same as the one stored in the Level 1b product except that it is re-ordered in increasing wavelength (in the spectral overlap between the spectrometers A1 and A2). This quantity is the difference between the reference spectral grid and the actual spectral grid of the transmissions written in this product.

 

5.1.3.7 Near Real Time meteo product

 

All fields included in the NRT meteorological products are also included in the GOMOS level 2 products. Please refer to the previous chapter for the explanation of these fields.

 


Keywords: ESA European Space Agency - Agence spatiale europeenne, observation de la terre, earth observation, satellite remote sensing, teledetection, geophysique, altimetrie, radar, chimique atmospherique, geophysics, altimetry, radar, atmospheric chemistry