Data Set Overview
=================

The Composite Infrared Spectrometer is a dual-interferometer carried on the
Cassini spacecraft Remote Sensing Palette. Cassini was launched on 15
October 1997 and due to arrive at Saturn on 1 July 2004. En route to
Saturn, Cassini made a gravity-assist maneuver at Jupiter in December
2000, which allowed for a six-month Jupiter observing campaign. CIRS was
operated in full science mode for much of this six month period,
either as the 'prime' instrument controlling pointing, or as a
'ride-along' when another team was 'prime'. The CIRS Jovian dataset
forms the first part of the CIRS archive.

The Cassini Prime Mission after SOI (Saturn Orbit Insertion) is the
first four years of tour, from July 2004 to July 2008. After that
time, if approved and technically possible, an extended mission period
will begin. During the first four years of Prime Mission, Cassini will
undergo 76 orbits (revs) of Saturn, known as 0, A, B, C, 3 ... 74. The
original orbits 1 and 2 were replaced with A, B and C when the mission
was redesigned in order to accommodate radio link issues with the
Huygens probe, discovered after launch. The probe delivery to Titan is
scheduled for rev C. There are 44 targeted (i.e. close) fly-bys of
Titan planned during the first 76 revs.

See MISSION.CAT file for more information on the Cassini mission.

Science Objectives and Observation Strategy
-------------------------------------------

The Cassini/Huygens mission to
the Saturnian system is designed to investigate the following targets:

1. Saturn
2. Rings
3. Titan
4. Other ('icy') satellites.

The overall mission may also be broken in physical target types:

1. Atmospheres (Saturn and Titan)
2. Surfaces (Titan, and other satellites).
3. Magnetosphere: fields, particles and solar wind interaction (Sun-Saturn)
4. Orbital mechanics (rings, satellites - especially Hyperion).

CIRS contributes to the investigation of many of these areas:

SATURN

Sensing of tropospheric and stratospheric temperatures and composition.
This includes abundances of the major and minor species, the hunt for new
gaseous species, isotope ratios for major species, and
dynamics. Allocating time for Saturn observations was the
responsibility of the Saturn TWT (Target Working Team), which was in
charge of much of the periapse parts of each orbit (rev). There was
also input from the Atmospheres Working Group (AWG) which dealt with
high-level science recommendations.
s
CIRS Saturn atmospheric science goals are met with observations at a
variety of inclinations, range from Saturn and, in some cases, in
conjunction with other teams' observations.  The standard CIRS requests
include:

Name    Range   Details

COMPSIT >60Rs   Composition sit and stare at 0.5 cm-1 spectral resolution
                using FP1. Uses all 3 focal planes. Search for oxygen
                compounds (CO2 and H2O) and new molecules.

MIRMAP  25-40Rs Mid-IR maps, centered at a latitude, with the long axis of
                our arrays (Z) pointed toward the pole of Saturn to allow
                the planet's rotation to map out a latitude band. 3 cm-1
                resolution.

FIRMAP  15-25Rs Far-IR maps, constructed by slewing equatorward over a
                hemisphere.  Slews overlap and repeat to map out the entire
                hemisphere. 15 cm-1 spectral resolution.

FTRACK <10Rs    Campaign with VIMS and ISS to track features across the disk
                from limb to limb, including limb sounding.  FP1 nadir
                sounding at 3 cm-1, FP3/4 limb sounding at 15 cm-1.
                The goal is to study vertical structure and poleward
                heat transport.

LIMBMAP <6Rs    Vertical sounding at multiple latitudes on the limb at 15
                cm-1.

LIMBINT <6Rs    Long integrations for vertical composition at one location on
                the limb. Resolution = 0.5 cm-1 or 1 cm-1 (TBD).

NADIROCC<10Rs   Helium abundance measurements taken by following the Radio
                Science Team's occultation points across the disk, 3 cm-1 res.

OCCLIMB <6Rs    Independent verification of vertical temperature profile at
                Radio Science's occultation latitudes.

REGMAP <10 Rs  Regional mapping and/or composition feature tracks, to fill
COMPFT/         in coverage not obtained by the other observations (0.5 or 3
                cm-1)

TEMPSIT         Early temperature map. Res=3 cm-1. Sit on CML, FP3/4 N-S, let
                planet rotate to cover all longitudes. Arrays cover Southern
                Hemisphere (north blocked by rings).

MIRCMPSIT        Early composition sit and stare. Res = 0.5 cm-1. Sit on
                CML, FP3/4 N-S, Arrays cover Southern Hemisphere
                (north blocked by rings).

FIRCMPSIT        Early far-IR composition. Res = 0.5 cm-1. FP1 centered
                on South Pole and on mid-Southern latitudes.


RINGS

Measurement of the infrared spectrum at varying phases, leading to
conclusions about ring particle size, shape, composition, distribution
and dynamics. Allocation of time for rings observing was carried out
by the Rings TWT, in conjunction with the Rings Working Group (RWG).

Ring observations are made as a function of ring opening angle, or
spacecraft elevation:

* Faint Ring Long Integrations.   The low optical depths of the
faint D, E, F and G rings will pose particular observing challenges
for CIRS. These rings are best viewed edge-on because this geometry
enhances the instrument fill factor. Low spectral resolution of 15
cm-1 with FP1 provides the best signal-to-noise and should be
sufficient for detecting the variations of emissivity with wavelength,
which is our primary measurement goal.  From close range (~ 10 RS) and
small opening angle, the FP1 filling factor will approach 1% when
pointed at the F rings ansa. Integrations of ~ 10 minutes should yield
usable signals. However, because the F ring is so clumpy, it needs to
be sampled at many longitudes before a truly representative spectrum
can be obtained. Observations will consist of alternating between both
ring ansas every ~30 minutes to achieve the most complete rotational
coverage of this ring. The E ring will be observed by pointing FP1
near the orbit of Enceladus, where the long edge-on line of sight
through the ring maximizes the fill factor. However, this fill factor
will still remain quite low, ~ 10-4, so , detecting the E ring will
require many, perhaps 100 or more, hours of integration. On the other
hand, because the ring is so thick vertically, the observing range can
be quite large (3040 RS). More observing time is available then during
these apoapse periods of the tour. The VIMS and UVIS instruments will
also require substantial integration on this ring, so E ring
observations will be cooperative activities between all of Cassini's
optical remote sensing instruments. Unfortunately, the best possible
fill factors for the remaining rings, D and G, are still lower than
for Ring E. It is unlikely that either will be detected with CIRS.

* Composition Integrations.  CIRS will determine with unique
accuracy the ring spectrum between 50 and 1000 micron. As intimately
mixed contaminants significantly influence this part of the spectrum,
mixtures derived from the visible and near-infrared spectra will be
tested against this new spectrum. Spectra of the three main rings over
the full CIRS wavelength range will be obtained to determine possible
radial variations in the bulk composition.  Two types of observations
will be made: high spectral resolution (0.5 cm-1) FP3 emission
measurements of the A, B and C rings, and high spectral resolution
transmission measurements of the rings with the rings against Saturn.
The former can be obtained from large ranges 2040 Rs because of FP3s
fine spatial resolution; long integrations of 1020 hours will be
obtained on representative locations in each ring. The transmission
measurements will be made from 20 Rs at relatively low ring opening
angles.  This will allow a search for absorption features in the A and
C rings, and the Cassini division.   The same region of Saturn will be
observed in at a similar spatial resolution when the rings are not
present, to establish the background.  The transmission spectra will
be obtained over a series of emission angles.

* Stellar Occultations. A handful of stellar occultations are
observed by CIRS to directly obtain the ring opacity in the infrared.
Only a limited number targets are observable by CIRS, including CW Leo
and Eta Carinae.  Eta Carinae occultations are only observable during
the final month of the tour.  Occultations are observed in one FP3
pixel (CW Leo) or one FP4 pixel (Eta Carinae) at 15 cm-1 spectral
resolution.

* Radial scans. These scans are typically executed between 5 and
20 Rs over a range of spacecraft inclinations, from low (5 deg) to
highest possible inclination (75 deg), radial mapping (FP1, FP3) of the
rings, on both lit and unlit sides, over a range of spacecraft
elevations, inclinations, local times and phase angles, is performed
to obtain broadband radiometric measurements of the total flux in the
CIRS wavelength range. Sets of observations are obtained in each of
the inclined orbit intervals to map the temperature variation in the
rings with changing solar illumination.  Two types of scans are
planned. So called temperature scans will consist of spectra at 15
cm-1 spectral resolution of the lit and unlit sides of the rings at
many incidence and emission angles and provide prime information on
the ring thermal gradient as a function of radial distance to
Saturn. Submillimeter scans will be made of spectra at 1 cm-1 spectral
resolution of the lit and unlit sides of the rings to map the thermal
characteristics and composition of the ring particles out to 1 mm.

* Azimuthal scans. These observations are executed between 5 and
20 Rs at spacecraft inclinations greater than 20 deg.  They will be used
to study both the surface properties, the vertical dynamics and the
spin of ring particles. Observations of the cooling and heating of the
ring particles entering and emerging from the planetary shadow are
planned to derive particle thermal inertias for all three main
rings. It will make measurements at moderate radial resolution
(typically 1000 km) across the shadow boundaries at low spectral
resolution (15 cm-1) with the FP1 field of view. To constrain the
vertical dynamics of ring particles, the temperatures of the main
rings will be measured by CIRS along the ring azimuth of the main
rings, from the exit of the shadow (morning) to the evening ansa, both
on the unlit and unlit faces. This unique experiment will be realized
with spectra at low spectral resolution (15 cm-1). Spins create both
an azimuthal asymmetry in the ring temperature and a dependence of the
temperature with the emission angle, due to day/night
contrast. Circumferential scans at a variety of phase and emission
angles will be executed to detect azimuthal asymmetries and the
anisotropy in the ring particle emission function which are both
function of particles spin and thermal inertia. Occasionally, when
observing time is highly disputed, long azimuthal scans (8-to-20 hours
long depending on geometry and face) will be replaced by a series of
radial scans at different azimuths.

TITAN

Sensing of tropospheric and stratospheric temperatures and composition.
This includes abundances of the major and minor species, the hunt for new
gaseous species, isotope ratios for major species, and dynamics.
CIRS may also be able to sense the surface near 600 cm-1. Allocation
of time for Titan observations was primarily done in the TOST group
(Titan Orbiter Science Team), in conjunction with recommendations from
the AWG.

CIRS can achieve different science goals at different distances from
Titan. Typically, CIRS makes the following requests (symmetric about
closest approach):

+ 0 to +10 mins   HIRES surface mapping (e.g. slew over south pole).

+10 to +45 mins   FIRLMBT - radial limb scans with FP1 to derive
                temperatures in the 8--100 mbar region via the N2-N2
                collision-induced absorption between 20--100 cm-1.

+45 to +75 mins   FIRLMBAER - radial limb scans with FP1 to measure/
                characterize particulate and condensate distributions,
                abundances and properties.

+75 to +135 mins  FIRLMBINT - integrate at two altitudes on the limb
                with FP1 to search for signals of CO, H2O and new
                species.

+2:25 hrs to +5 hrs FIRNADMAP - slow scan north-south or east-west on
                the disk to sound tropospheric temperatures at 40--200
                mbar, via the N2-N2 absorption,
        OR,
                slow scans at constant emission angle on the disk,
                to retrieve surface temperatures in the presence of
                aerosols around 520 cm-1.

+5 to +9 hrs    MIRLMBMAP - map 1/4 limb using the FP3 and FP4 arrays,
                to infer stratospheric temperatures via the 1304 cm-1
                band of CH4. The arrays are placed perpendicular to
                the limb at two altitudes, chosen to provide
                overlapping coverage of the altitude range 150 to 420
                km. The arrays are used in blink (ODD-EVEN)
                mode. After mapping both altitudes, the arrays are
                stepped 5 degrees in latitudes for the next step.
        OR,
                MIRLMBINT - as in MIRLMBMAP, except that only a
                single latitude is covered, at two over-lapping
                altitudes for 2 hrs in each position. To search for
                and measure new species in the mid-IR: methyl, benzene
                etc

+9 to +13 hrs   FIRNADCMP - integrate on the disk at emission angle
                approximately 60 degrees with the FP1 detector, in
                order to measure spatial abundance distribution of
                weak species and search for new species in the far-IR.

+13 to +22 hrs  MIDIRTMAP - scan the entire visible disk with the
                FP3/FP4 arrays perpendicular to the scan direction
                ('push-broom'), to measure stratospheric temperatures
                via the CH4 v4 band. Used for later dynamical
                analysis, for winds, waves etc.

+22 to +48 hrs  COMPMAP or TEMPMAP - map a meridian across the planet
                either E-W or N-S, using the Fp3/FP4 arrays in two
                positions longwise. To search for new species, and/or
                monitor temperatures.

ICY SATELLITES

Surface mapping in the IR, providing information on the surface composition,
morphology, and age. Close passes of icy satellites resulted in time
allocated to the SOST (Satellites Orbiter Science Team), which divided
the time between teams. There was also input from the Surfaces Working
Group (SWG).

During the Cassini tour, there are eight flybys of the classical icy
satellites targeted at 1000 km or less, as well as a number of
'Voyager-class' (less than 300,000 km) encounters.  There are also
flybys of several much smaller satellites, such as Janus and
Epimetheus, at various distances.  The dimensions of these objects
range from less than 100 km to as much as 1530 km (for Rhea).
Consequently, it is most useful to discuss CIRS observations in terms
of the angular diameter of the object, rather than its distance from
the spacecraft.

Normally the spacecraft orientation is controlled by momentum wheels,
which provide pointing accuracy and precision of ~2 mrad and ~0.04
mrad, respectively.  Consequently, the full spatial resolution of the
FP3 and FP4  pixels (0.3 mrad) cannot be utilized with reasonable
confidence until the target exceeds 1 mrad in diameter.  At that
point, the separation of focal planes 3 and 4 (0.9 mrad), and the
large size of FP1 (nominally 3.9 mrad) offer reasonable assurance of
obtaining useful data in either FP1 or at least one of FP3 and FP4.
As with the planet and the rings, the order of magnitude difference in
FOV scale between FP1 and the other focal planes plays heavily in the
design of the icy satellite observations.  Approaching from a
distance, CIRS observations might proceed roughly in the following
order.

        FP1INT, FP34INT:  Compositional integrations (typically
        performed at ranges where the target AD is 1-3 mrad).  Because
        spectral features of solid materials are generally broader
        than those of molecular lines, these observations are made at
        1-3 cm-1 resolution.  Conducted in staring mode, using FP1 and
        center detectors of FP3 and FP4, these are concentrated in
        geometries with low phase angles, so as to view the icy
        satellite surfaces at their warmest and maximize the
        signal-to-noise ratio.  The steep rolloff of intensity with
        increasing wavenumber requires increasing integration time to
        extend the spectrum.  For a 100K surface and a spectral
        resolution of 3 cm-1, a SNR of 10 at 600 cm-1 can be obtained
        with 1 minute of integration, but extending the spectrum to
        800 cm-1 with the same SNR requires integration of 100 minutes
        Therefore, extending the spectrum to the highest wavenumbers
        will involve coadding spectra over the entire tour.

        SECLN, SECLX:  Satellite eclipse entry (N) or exit (X)
        observations (AD>1 mrad, phase angle phi <100 degrees,
        spectral resolution delv = 15.5 cm-1; a
        typical eclipse duration is ~2 hours).  The high sensitivity
        of FP1 permits observations of objects even when the FOV is
        incompletely filled, or when the phase angle is moderately
        high.  At low phase angles, even FP3 can be used to follow the
        initial portion of the cooling curve for relatively dark
        objects; this allows observations with body-centric
        spatial resolution better than 10o to be made when the
        apparent target size exceeds ~3 mrad.  Eclipses provide the
        thermal inertia of the upper mm or so of the surface, as well
        as estimates of surface coverage by relatively large fragments
        of consolidated material.

        FP1FAZ:  Phase/longitude (diurnal cycle) coverage (AD>1 mrad,
        delv=15.5 cm-1, 10-20 minutes). Focal plane 1 will be utilized to
        determine the disk-averaged temperature of the satellites.
        From the resulting diurnal behavior, mean thermal inertias in
        the upper cm or so of the surface will be derived.

        FP34FAZ, FP34MAP, FP34REG:  Global thermal inertia mapping
        and/or hot spot monitoring (AD>3 mrad, delv=15.5 cm-1).  Maps
        are made by slewing and rastering FP3 and FP4 across the disk
        at rates not to exceed that for Nyquist sampling (16 microrad/sec
        in blink mode); observation durations will typically be 10-30
        minutes.  These maps will be successful for varying portions
        of a satellite, depending on its albedo and thermal inertia.
        The exception is Enceladus, which is so cold that even the
        subsolar regions will be barely detectable in an individual
        measurement.  In this case, mapping will serve to monitor the
        satellite for ongoing endogenic activity; for example, active
        sources at or above the NH3.H2O eutectic temperature will
        be easily observed if they fill more than a few percent of an
        FP3 or FP4 pixel.

        FPGREEN:  Search for a solid state greenhouse (AD>10 mrad,
        phi <40 degrees, delv =15.5 cm-1, 15 minutes).
        The decreasing absorption coefficient of water
        ice with decreasing wavenumber in the
        far-IR permits detection of radiation from increasingly far
        below the surface, reaching as deep as 1 cm at 10 cm-1. Slow
        east-west slews using FP1 allow following the daily
        penetration of the thermal wave into the regolith.

        FP1POLE:  Polar night (annual cycle) coverage (AD>10 mrad,
        delv=15.5 cm-1, 15 minutes).  FP1 observations of the dark winter
        polar region from high latitude permit determination of the
        seasonal cooling curve.  This enables an estimate of thermal
        inertia in the upper tens of cm of the polar regolith.

        FP1DAYMAP, FP1DRKMAP:  Hemispheric FP1 mapping near closest
        approach (AD>10 mrad, delv=15.5 cm-1).  The high sensitivity of
        FP1 permits rapid mapping of satellite disks near closest
        approach, where time is at a premium.  Nyquist-sampled maps at
        low spectral resolution (slew rate 400 microrad/sec) will
        typically take 10-30 minutes.  These permit identification of
        minor thermal anomalies, even on the night hemisphere.

Other TWT and WGs

The cross-discipline TWT (XD TWT) allocated time near apoapse on some
revs. The magnetospheres TWT allocated critical time for magnetospheric
observations, such as time near critical transition boundaries.


Instrument Overview
-------------------

CIRS consists of two interferometers; a far-IR polarizing interferometer
sensitive from 10-600 cm-1, and a mid-IR interferometer sensitive from 600
-1400 cm-1. The far-IR radiation is sensed by a large (~4 mrad diameter)
thermopile detector (actually two, one for reflected and one for
transmitted beam at the polarizer/analyzer), known as FP1. The mid-IR
radiation falls on two 1X10 arrays, known as focal planes 3 & 4. FP3
is an array of photoconductive detectors sensitive from 600-1100 cm-1,
and FP4 is an array of photovoltaic detectors sensitive from 1100-1400 cm-1.
FP2 was dropped from the original design.

Each of the 1X10 arrays has a 5-channel amplifier/signal processor, meaning
that only 5 of the 10 detectors in each array can be used at a time. The
maximum number of simultaneous interferograms from CIRS is therefore 11
(1 from FP1, 5 from FP4, 5 from FP4). However, there are observing modes
in which some of the focal planes are not used, leading to a lesser number
of active channels.

Almost the entire instrument is thermostated to 170 K, including the
primary and secondary mirrors, internal optics, scan mechanisms, and the FP1
focal plane assembly. The only exception is the FP3 and FP4 arrays (the
mid -IR FPA), which are cooled by a passive radiator to between 74 and 85K.

See the INST.CAT file for full details.

Calibration Overview
--------------------

Wavelength calibration is achieved by use of a reference diode laser
interferometer which uses the same scan mechanism as the IR
interferometer. Careful attention must be paid to the laser mode and
voltage which may drift/mode jump.

Radiometric calibration in the far-IR (FP1) is simplified due to the
fact that there is only one temperature inside the
instrument. Therefore, use of a deep space (2.7 K) reference target is
sufficient to calibrate radiance.

The fact that the mid-IR has two temperatures to contend with means
that a second reference target is required. Therefore, in the mid-IR
optical path, a shutter may be lowered for calibration purposes, which
gives a 170K reference body.

Calibration must be done using complex number algebra and Fourier
transforms, to correctly account for phase changes which occur in the
beamsplitter. See the DATASIS.PDF for details.

Parameters
----------

The major parameters when observing with CIRS are as follows:

1) mid-IR shutter (open or closed). This is used for radiometric
calibration in the mid-IR, which requires two thermal reference
bodies (space and shutter).

2) mid-IR focal plane temperature set point (74-85 K). This was
usually set to the lowest level in the range at which the radiator
could cope with heating from the sun and/or Saturn/rings, and
therefore, achieve stability of FPA temperature.

3) FP3 pixel mode (ODD, EVEN, CENTER, PAIRS). Odd mode: detectors
1,3,5,7,9 active. Even mode: detectors 2,4,6,8,10 active. Center mode:
detectors 3,4,5,6,7 active. Pairs mode: 1&2, 3&4, 5&6, 7&8, 9&10
signals combined and amplified.

4) FP4 pixel mode (ODD, EVEN, CENTER, PAIRS). As FP3 modes, except
that detectors 4,5,6,7,8 used in CENTER mode, to spatially match the
detectors of FP3 which are numbered in reverse physical order.

5) co-add mode (COADD or NO-COADD). In co-add mode, two successive
scans are added together in the on-board electronics (buffers), to
reduce the data rate by half.

Data
----

This dataset is composed of CIRS Time Sequential Data Records,
CIRS Spectral Image Cubes and related calibration software.

READOUT, DOWNLINK AND DECOMPRESSION

CIRS data is normally read out at a rate of 4 kilobits per second
(kbs) and 8kbit packets are transferred via the Bus Interface Unit
(BIU) to the spacecraft Command Data Subsystem (CDS) every 2 seconds.
CIRS has many ways of reducing data rate, such as co-adding IFMs
(reduces rate by factor 2), dropping FP3 or FP4 readout (reduces by
45%), dropping FP3 and FP4 readout (reduces by 91%), or going to
housekeeping only mode.

Data packets are compressed in the instrument electronics, packetized
on the spacecraft, transmitted to the Earth via the Deep Space Network
(DSN) and then decompressed by software on the ground. At this stage,
the data goes into a processing pipeline, which organizes the science
data into binary tables and interpolates housekeeping
records. Pointing and geometry tables are also produced when the NAIF
data becomes available.

VANILLA PROGRAM FOR READING TABLES

The data is in the form of binary tables, also known as the
Vanilla database. Vanilla is a simple database access tool,
also included on each volume as both source and executable.
Note however that the Vanilla software is NOT REQUIRED in order
to read the binary data, but it may simplify the task. Notably,
Vanilla links together all binary files into a huge, `logical'
table, which can be searched based on key fields.

However, the user may read the binary data directly, using any
programming language desired. If this approach is taken it is
extremely important to pay attention to binary field widths:
field widths vary, and typically trying to read an entire record
into a data structure will fail: due to assumed padding of fields.
The user must read the data in field by field, specifying width.
Field widths are described in the *.FMT files, which are placed
in each data directory.

Some tables have both fixed-length and variable-length fields.
In this case, the variable length part of the record is stored
separately, in a file with the same name but an extension .VAR,
instead of .DAT. The third type is file extension is .LBL, which
is the PDS detached label.

DATA LEVELS

The level-0 tables (base level) are contained in /DATA/TSDR/UNCALIBR.
These are described in detail in the document DATASIS.PDF
(DATASIS.TEX). These table types (IFGM, OBS, IHSK, FRV) are essentially
the raw-level information which came from the instrument: the
only process applied was unpacking (uncompressing) the instrument
packets and, in the case of the housekeeping data, interpolating
the 64-sec interval housekeeping data onto the science scan intervals.

Original housekeeping data is stored in /DATA/TSDR/HSK_DATA, which is read
out at 64-sec intervals when the instrument is either ON or in SLEEP mode.

Pointing and geometry information is stored in /DATA/TSDR/NAV_DATA.
These tables are described in detail in the DATASIS document, and
also in the FMT format files in the NAV_DATA area. Primarily, these
store the spacecraft attitude and position with respect to major bodies
(GEO file type), the position of the 11 detectors (9 Q-points,
or fiducial reference marks) on all bodies in the FOV (POI file
type), and the list of bodies seen in each detector (TAR type).

Pointing and geometry information is derived from the spacecraft
re-constructed attitude ('C') and trajectory ('SPK') kernels, using
NAIF (Navigation and Ancillary Information Facility) toolkit routines
supplied by JPL.

Apodized calibrated spectra are stored in /DATA/TSDR/APODSPEC.
Apodization is the process of mathematical filtering or smoothing
which removes `ringing' effects from finite-width mathematical FFTs,
widening the instrument line shape in the process. Apodization will
normally use the Hamming function window, but other types may be used
and the type will be given in the tables.

CUBES

Spectral Image Cubes are generated by systematic processing
of the calibrated spectra. These data products are described
in detail in the document CUBESIS.PDF.

Cubes are generated using two types of projections : equirectangular
and point-perspective. These are stored in DATA/CUBE/EQUIRECTANGULAR,
 DATA/CUBE/POINT_PERSPECTIVE and DATA/CUBE/RING_POLAR.

This volume represents a first attempt at calibrating the CIRS Jupiter
dataset. Certain calibration problems are known to exist, in
particular, thermal drifts in the ATMOS02B map noticeable at the 1-2 K
level, which are seen as 'striping'. Similar drifts are likely to be
present in other areas. Hence, the absolute radiometric accuracy is
unlikely to be better than 1%. If a way is found to adequately remove
these effects, it is envisaged that a second edition of the data would
be produced. The spectral calibration is thought to be accurate to
approximately 0.1 cm-1.

    NOTE ON INTERFERENCES

CIRS interferogram data suffers from a number of external interferences,
especially:

- an 8 Hz spike pattern due to spacecraft communications and also to the
  onboard numerical filtering.

- a 1/2 Hz spike pattern due to the Bus Interface Unit, transfer of data.

- a sine wave of variable frequency which appears correlated with the
  electronics board temperature.

- scan speed fluctuations which have been traced to two mechanical vibrations
  on the spacecraft: (a) the MIMI LEMMS actuator (b) the reaction wheels used
  to turn the spacecraft.

- a 1 Hz spike pattern from the analog multiplexor data readouts.

- an 8.3 Hz spike pattern from an unknown source.

These various effects, plus the onboard and on-ground processing done
to mitigate them, are described in more detail in the
cirs_interferences.pdf, CIRS-USER-GUIDE.PDF, and CirsNoise.pdf documents
found in the DOCUMENT directory.

    PDS VERSION 2.0

The entire CIRS dataset has been re-delivered as Version 2.0
(DATA_SET_ID = CO-S-CIRS-2/3/4-TSDR-V2.0), which brings all previous
volumes up to the same calibration level.

In this release, the AREA8HZSPIKE field in the UNCALIBR/DIAG table was
removed and the following fields added:

   NAME = RATIO_SINE_8HZSPIKE
   NAME = ZPD_POS
   NAME = ZPDPEAK
   NAME = RAWPOWER
   NAME = DELTA_BIURTI
   NAME = FIFM_STD_DEV
   NAME = FIFM_ID
   NAME = TZPD
   NAME = TAMP
   NAME = ZPDENV
   NAME = AMPENV
   NAME = DCLEVEL

There were some slight changes made to the calibration algorithm used to
identify blocks of deep space data. In those cases where a deep space data
gap of more than an hour was found, a new minor block was created. Also
modified, the deep space search window was increased from 4 hours to 8
hours, and the maximum number of deep space interferograms allowed in a
block was increased to 5000.

    PDS VERSION 3.1

The entire CIRS dataset was re-delivered as Version 3.1 (i.e., DATA_SET_ID
= CO-S-CIRS-2/3/4-TSDR-V3.1). Following is a list of changes for each of
the vanilla tables affected by this release. Descriptions for new fields
can be found in the .FMT file for that table or in DATASIS.PDF.

Table HSK_DATA/HSK:

   Field DECSPOSN changed from LSB_UNSIGNED_INTEGER (2 bytes) to PC_REAL
   (8 bytes).

   Field PHOTODIODE changed from LSB_UNSIGNED_INTEGER (2 bytes) to PC_REAL
   (8 bytes).

   The second IDSCALIB field has been renamed to IDSCALIB2.

Table UNCALIBR/DIAG:

   New fields:
   NAME = RWA1_MIN
   NAME = RWA1_MAX
   NAME = RWA2_MIN
   NAME = RWA2_MAX
   NAME = RWA3_MIN
   NAME = RWA3_MAX
   NAME = RWA4_MIN
   NAME = RWA4_MAX
   NAME = RWA_NOISE_FLAG

Table UNCALIBR/OBS:

   Eight new fields were added:
   NAME = SCET_MSEC
   NAME = COMPUTED_RTI
   NAME = CONSECUTIVE_NULL_SCANS
   NAME = IDS_MUX
   NAME = TCM_MUX
   NAME = PACKETIZATION_STATUS
   NAME = MSEC_SINCE_RTI
   NAME = FSV

Table NAV_DATA/GEO:

   Changed TIME_ZPD from LSB_INTEGER to PC_REAL. It is now a time based on
   SCET, OBS.SCET_MSEC, and DIAG.TZPD.

   The field SCET_FRACTIONAL_SECONDS has been removed.

   The following fields were added:
   NAME = BODY_SUB_SPACECRAFT_LATITUDE_PC
   NAME = BODY_SUB_SOLAR_LATITUDE_PC
   NAME = PRIMARY_SUB_SPACECRAFT_LATITUDE_PC
   NAME = PRIMARY_SUB_SOLAR_LATITUDE_PC

Table NAV_DATA/POI:

   Changed TIME_ZPD from LSB_INTEGER to PC_REAL. It is now a time based on
   SCET, OBS.SCET_MSEC, and DIAG.TZPD.

   The following fields were added:
   NAME = ZLIMB
   NAME = SPACECRAFT_BODY_FIXED
   NAME = FOV_BODY_FIXED
   NAME = Z
   NAME = LATITUDE_ZPD_PC
   NAME = LATITUDE_END_PC

Table NAV_DATA/RIN:
   Changed TIME_ZPD from LSB_INTEGER to PC_REAL. It is now a time based on
   SCET, OBS.SCET_MSEC, and DIAG.TZPD.

   The fields RING_LONGITUDE_ZPD and RING_SOLAR_ZENITH are now defined for
   all Q points.

In addition to the above, this release includes an algorithmic change.
For interferograms collected using CIRS Flight Software Version 6 (executed
in July of 2010), non-co-added data will have the 1/2 Hz noise spikes
suppressed while no noise spike suppression will be done on co-added data.

    PDS VERSION 3.2

The entire CIRS dataset was re-delivered as Version 3.2 (i.e., DATA_SET_ID
= CO-S-CIRS-2/3/4-TSDR-V3.2). Following is a list of changes for each of
the vanilla tables affected by this release. Descriptions for new fields
can be found in the .FMT file for that table or in DATASIS.PDF.

Table HSK_DATA/HSK:

   HSK.HSECOOLTEM: This field previously contained negative temperatures
   for five scans in 2005 and 2006. Each of those scans corresponds to an
   analog to digital conversion error when trying to sample the MUX channel
   which sets an error flag in housekeeping in the following packet. This was
   not being picked up in the previous version. Now the flag is checked and
   if set, HSK.HSECOOLTEM is set to -1.

Table UNCALIBR/OBS:

   OBS.SCLK:  Changed from LSB_UNSIGNED_INTEGER to PC_REAL. It represents
   the time equivalent of OBS.SCET + OBS.SCET_MSEC / 1000.

   OBS.SCET + OBS.SCET_MSEC / 1000. now represents the time of the first raw
   data sample. Because fractional seconds are now included in the SCLK to
   SCET conversion, going from v3.1 to v3.2 sometimes increases the SCET value
   by 1 second. This occurs when the fractional seconds, computed_rti, and
   msec_since_rti are combined to get a SCET_MSEC value greater than 999.

   NAIF SPICE kernels are now used to convert from SCLK to SCET in the OBS
   table. (Previously an epoch specified in the SFDU header was used.) This
   change in methods may produce a correction in the OBS.SCET value when
   calculations switch from using predict SCLK kernels to reconstructed
   kernels.

   OBS.SCAN_FLYBACK_MSEC: Added. Represents the time (in seconds) from the
   first raw data sample to the start of the mirror flyback. Implemented
   on board the instrument as of FSV 6.0.1. Before FSV 6.0.1, this field
   will be set to 0.

Table UNCALIBR/IHSK:

   IHSK.HSECOOLTEM: The previous version allowed the use of negative
   HSK.HSECOOLTEM values when calculating the interpolated value
   IHSK.HSECOOLTEM. This version only uses non-negative values of
   HSK.HSECOOLTEM for the interpolation.

Table NAV_DATA/GEO:

   GEO.TIME_ZPD: Calculation change. It now uses OBS.SCET, OBS.SCET_MSEC,
   and DIAG.TZPD, along with the focal plane dependent time delays associated
   with the numerical filter to find a much more accurate ZPD time (in the
   SCET time system). Time shifts of up to 3 seconds occur because of the
   change.

   GEO.SCLK: Changed from LSB_INTEGER (4 bytes) to PC_REAL (8 bytes).


Table NAV_DATA/POI:

   POI.TIME_ZPD: Calculation change. It now uses OBS.SCET, OBS.SCET_MSEC,
   and DIAG.TZPD, along with the focal plane dependent time delays associated
   with the numerical filter to find a much more accurate ZPD time (in the
   SCET time system). Time shifts of up to 3 seconds occur because of the
   change.

   POI.SMEAR: Calculation change. The condition that the boresight hit the
   target in order to compute the SMEAR field has been removed. It will
   still compute the same value as in the previous version if the boresight
   hits the target, but the new version always computes the SMEAR. When
   it doesn't hit the target it now uses the ray periapsis rather than
   the intersection point.

   POI.Z: A bug in the code which used the wrong indices to calculate this
   field has been corrected.

Table NAV_DATA/RIN:

   RIN.TIME_ZPD: Calculation change. It now uses OBS.SCET, OBS.SCET_MSEC,
   and DIAG.TZPD, along with the focal plane dependent time delays associated
   with the numerical filter to find a much more accurate ZPD time (in the
   SCET time system). Time shifts of up to 3 seconds occur because of the
   change.

Table NAV_DATA/TAR:

   Three new targets have been added to the TAR table: Daphnis,
   Methone, and Pallene. These correspond to three new fields TAR.DAPHNIS,
   TAR.METHONE, and TAR.PALLENE. New bit fields in TAR.FOV_TARGETS have also
   been defined with bits 2^26, 2^27, and 2^28 represented by Methone,
   Daphnis, and Pallene, respectively.

Table APODSPEC/ISPM:

   The selection criteria fields used in the compilation of calibration
   blocks are now detector, rti, coadd, shutter, laser mode (1,2,3,!=0),
   fpa setpoint, noise (!=1), fp1dettem, hsecooltem, fpatem,
   firpolriztem, npts, truezpd, rawpower, and flight software version.
   Previously, only detector, rti, coadd, fpa setpoint, noise (!=1) and
   optical sense mode were used.

   The following fields were added:

        NAME                = CALIB_SCORE1
        NAME                = CALIB_SCORE2
        NAME                = CALIB_SCORE3
        NAME                = CALIB_SCORE4
        NAME                = CTZPD1
        NAME                = CTZPD2
        NAME                = DIST_DS_TZPD1
        NAME                = DIST_DS_TZPD2
        NAME                = DIST_DSSH_TZPD1
        NAME                = DIST_DSSH_TZPD2
        NAME                = DIST_SH_TZPD1
        NAME                = DIST_SH_TZPD2
        NAME                = DS_TDET
        NAME                = DS_TINSTR
        NAME                = DS_TZPD1
        NAME                = DS_TZPD2
        NAME                = FPASET
        NAME                = LASER_WL
        NAME                = PHASE_SHIFT
        NAME                = HASE_SHIFT_ERROR
        NAME                = PHASE_SHIFT_FLAG
        NAME                = SH_TDET
        NAME                = SH_TINSTR
        NAME                = SH_TZPD1
        NAME                = SH_TZPD2
        NAME                = TDET

Newly included in the archive are the spectral image cube dataset (V1.0)
and the Titan atmosphere temperature, aerosol, and ice profiles.

    PDS VERSION 4.0

The entire CIRS dataset was re-delivered as Version 4.0 (i.e., DATA_SET_ID
= CO-S-CIRS-2/3/4-TSDR-V4.0). Following is a list of changes for each of
the vanilla tables and algorithms affected by this release. Descriptions
for new fields can be found in the .FMT file for that table or in
DATASIS.PDF.

The FP1 electrical interference spike suppression algorithm has been
improved for 0.5 Hz and 8 Hz de-spiking, and 1 Hz de-spiking and
sine wave suppression have been added.

A new algorithm for detecting and characterizing velocity variations in
interferograms for both pre- and post-FSW 6.0.0 has been added.
The algorithm examines the values of MECH_OUT_OF_PHASE, RWA_NOISE_FLAG, and
NPTS to establish two levels of velocity variation disturbance: NOISE = 1
('negligible' or 'mild' velocity variations) and NOISE = 4096 or 8192
('severe' velocity variations).

Made enhancements to the interferogram noise detector algorithm to more
effectively identify and reject interferograms afflicted with a wide range
of disturbances: spikes, drifts, DC level off, too few samples, too many
samples, incorrect ZPD position, and velocity variations.

The number of samples (FPTS) assigned to filtered interferograms (FIFMs)
for each focal plane and RTI has been refined. Every FIFM of a given focal
plane and RTI is assigned the same value of FPTS, regardless of its NPTS
value. In addition, 40 RTI interferograms are truncated to the values of
FPTS assigned to 39 RTI interferograms. Finally, all FP1 39 and 40 RTI
interferograms now have FPTS = 179 samples and are therefore symmetric
about the ZPD fringe at sample number = 89. These changes ensure that
all spectra of a given RTI have the same spectral resolution.

Improved the power spectrum (RAWPOWER) of the symmetrical
two-sided ZPD region of each interferogram. Focal plane 1 RAWPOWER is now
calculated after the DC baseline and deep-space shape function are subtracted
from the interferogram. The focal plane 3 RAWPOWER is calculated after
the DC baseline is removed.

There is a small difference in the way the laser value is being selected.
For any FRV[2] value that fits between a minimum and maximum value
(4 and 6.25, respectively), the laser mode is selected according to SCET.
If it is outside the limits, the laser mode is set to 0.

In previous versions, only scans with a noise value of 1 were written to
the FIFM tables. This version additionally allows all scans with noise
values of 4096 and 8192 to be written to the FIFM table. If the noise
value is other than 4096, 8192, or 1, the scan is written to the FIFM
table as having a length of 1 and a value of 0. This was done to allow
the scan to be included in the output of queries that include other tables.

New calibration algorithm:

   The calibration of CIRS data requires the calculation of reference
   interferograms or spectra at 2 known temperatures, commonly referred
   to as 'cold' and 'hot'. For our instrument, the cold data is acquired
   by pointing at the cold deep space, away from any known target. The
   hot data is acquired by closing the shutter and looking at the structure
   of our instrument. As part of the science observation designs, the
   instrument team included time periods for the acquisition of these 2
   types of reference data. Nominally, they would be included at least at
   the beginning and at the end of an observation, but if the observation
   is extended, the observation may have been designed to include additional
   calibration measurements during the observation.

   In CIRS volume deliveries before PDS version 4.0, to calibrate a given
   interferogram the algorithm was designed to find a contiguous block of
   reference data as near as possible in time to the interferogram and build
   a cold and a hot average. Because contiguity was a requirement, the
   averages would sometimes be based upon very few interferograms. For the
   current release, the algorithm was modified to use a much larger number
   of cold and hot calibration interferograms, removing the contiguity
   constraint and using up to 4,000 scans during a 3 month range about the
   acquisition time. The primary goal of increasing the number of scans in
   the hot and cold averages is to increase the signal-to-noise ratio.

   This release also includes a phase correction algorithm used to compensate
   sampling errors during acquisition of the interferograms. The
   interferogram sampling system is sensitive to external mechanical
   perturbations such as the spacecraft reaction wheel assembly.
   Under specific conditions, the opto-mechanical system controlling
   the moving mirror can trigger sampling away from the nominal
   position. The recorded interferogram is then time shifted, and
   its corresponding spectrum will have a baseline distortion.
   The phase correction algorithm predicts and identifies anomalous
   interferograms and applies a linear phase correction.

Table UNCALIBR/DIAG:

   The following fields were added:

        NAME                = PHASESINE
        NAME                = RATIO_FWHM_HALF_HZ
        NAME                = RATIO_WN_EIGHT_HZ
        NAME                = RATIO_FWHM_EIGHT_HZ
        NAME                = RATIO_EIGHT_HZ_SPIKE_AMPLITUDE
        NAME                = IFM_DC_LEVEL
        NAME                = STD_DEV_BEFORE_ZPD
        NAME                = STD_DEV_ZPD_REGION
        NAME                = STD_DEV_AFTER_ZPD
        NAME                = DELTA_RTI_DIFFERENCE
        NAME                = NOISE_DIAGNOSTIC
        NAME                = AMPSINEWAVE
        NAME                = AMP8HZSPIKE

   The RATIO_SINE_8HZSPIKE, DCLEVEL, ZPDPOS, AND ZPDPEAK fields were
   removed.

   The FIFM_ID field has been renamed as SPECTRUM_DIAGNOSTIC, however,
   FIFM_ID remains as an alias.

Geometry/Pointing algorithm:

   The size of Titan's penumbra was previously increased (from a radius
   of target+600KM to a radius of target+1K KM), however,
   it created a problem during some observations where the
   spacecraft was inside the newly increased penumbra and the
   code marked those scans as being deep-space, as opposed to
   being Titan scans, which confused the calibration code. Now,
   when the body_spacecraft_range is smaller than the
   penumbra (i.e., radius of target + padding) the code ensures
   those scans are cataloged as the target and not deep-space.

   The following observations were affected by the penumbra change:
   CIRS_106TI_REGMAP001_VIMS on March 26-27, 2009
   CIRS_106TI_REGMAP001_ISS on March 27, 2009
   CIRS_106TI_HIRES001_VIMS on March 27, 2009
   CIRS_138TI_GLOBMAP001_VIMS on September 24, 2010
   CIRS_169TI_REGMAP001_VIMS on July 24, 2012
   CIRS_169TI_HIRES001_VIMS on July 24, 2012
   CIRS_169TI_HIRES002_VIMS on July 24, 2012
   CIRS_169TI_FIRNADMAP002_PRIME on July 24-25, 2012
   CIRS_169TI_MIRLMBMAP002_PRIME on July 27, 2012

   Jupiter is now explicitly searched for in the field of view
   during the time period Cassini was in orbit about Saturn.

Table NAV_DATA/TAR:

   Two new targets have been added to the TAR table: Polydeuces and
   Aegaeon. These correspond to two new fields TAR.POLYDEUCES
   and TAR.AEGAEON. New bit fields in TAR.FOV_TARGETS have also
   been defined with bits 2^29 and 2^30 represented by Polydeuces
   and Aegaeon, respectively.

   Changed the FOV_TARGETS field from 33554432 (rings)
   to 0 (deep space) when the RING_SPACECRAFT_RANGE_ZPD[5]
   field is negative, which means the rings are behind
   the spacecraft.

New table, OISPM:

   During the calibration process, a filtering and noise detection
   algorithm is applied to the data to determine the quality of the
   interferograms. The algorithm looks for anomalies in the number
   of interferogram sample points (typically from scan mechanism
   velocity variations), dual zero phase difference (ZPD) peaks, and
   irregular sampling (see the NOISE field in DIAG table). While
   shortened interferograms or dual ZPD interferograms are not usable,
   interferograms with moderate sampling perturbations can produce
   calibrated spectra, but they may be of lower quality or noisier
   than spectra calibrated without anomalies. To ensure users of these
   lower quality spectra are aware of the data's limitations, a new
   table, OISPM, has been created and placed in a directory
   (DATA/TSDR/SUSPECT_SPECTRA) separate from the nominal spectra,
   forcing it to be queried independently of the ISPM tables.

   Also critical to the calibration process, the temperature of the
   focal plane assembly (IHSK.FPATEM) of the mid-infrared focal plane
   assembly (FPA) must be taken into account. This temperature is defined
   by the thermal balance between the 70K cooler and a heater placed near
   the detectors. During a CIRS observation, the temperature of the FPA is
   set to one or more of a discrete set of values where each value
   (set point) corresponds to a temperature range of 0.4 K about values
   spaced 0.75 K apart. If the instrument is not thermally stable or if
   the sun, Saturn, or Saturn's rings illuminate the FPA radiator, the
   FPA may not be operating at a nominal, steady, controlled temperature.
   Because the mid-infrared detector responsivities are dependent on the
   FPA temperature, in order to be properly calibrated, target interferograms
   should be matched with averages of deep space data acquired at the same
   FPA temperature (set point). If the target and deep space interferogram
   FPA temperatures cannot be matched, nearby deep space data (in time) at
   a different temperature will be used, with the resulting calibrated
   spectra placed in the OISPM table.

   There are several additional cases in which the calibration of an
   interferogram was determined to be questionable, causing its spectrum
   to be placed in the OISPM table:

      1. The calibration accuracy is dependent on the number of deep space
         interferograms averaged together. If too few are available, the
         noise increases to unacceptable levels, and the spectrum becomes
         suspect.

      2. During the mission, the laser state (DIAG.LASER) changed modes
         several times which caused some instrument recording modes to have
         little or no reference data. Calibrating target data with reference
         data from a similar, but different, instrument mode caused small
         mismatches in the wavelength scale.

      3. A subset of the data used a mode for which interferograms were
        co-added time-wise. This sometimes affected the noise spike pattern
        making it difficult for the de-spiking algorithm to do its job.
        This was exacerbated if no co-added reference data was available.