Data Set Overview : The dataset contains measurements from the infrared interferometer spectrometer and ancillary data. Each record of the dataset consists of a header and a spectral observation of Mars; the header contains pointing and other information on the geometry of the observation. The dataset is ordered by time as measured by the Data Acquisition System (DAS) time (1 DAS count is approximately 1.2 sec.). Two IRIS frames are completed every 35 DAS counts; every eighth IRIS frame is devoted to a calibration observation (alternately of interstellar space and an internal calibration blackbody), and is therefore absent from the calibrated dataset.  The spectral data consist of calibrated thermal emission spectra expressed as spectral radiances in Watt/cm**2/sr/cm**-1. The wavenumber corresponding to each spectral radiance value is not included in the dataset; the beginning wavenumber and the constant wavenumber increment are given, permitting easy calculation of the appropriate wavenumber for each radiance. The calibrated radiances have been obtained from the directly measured interferograms of Mars, along with the deep space and internal blackbody calibration observations. The interferograms are first symmetrized to correct for the fact that they are not sampled at zero path difference, and to remove the asymmetry due to residual dispersion in the beamsplitter and compensator. The symmetrized interferograms are then apodized using a Hamming function, and are cosine transformed. The responsivity obtained from the deep space and internal blackbody measurements is then used to obtain calibrated radiances.  Mariner 9 was placed in a 12 hour orbit about Mars at Lsubs:293, with data transmitted in real time during a 2 hour period around periapsis. The 12 hour period was selected to assure that the spacecraft was always visible from the single large receiving antenna at Goldstone, California once per day (the even numbered orbits). Data from alternate (odd numbered) orbits was received by a smaller antenna at Madrid, Spain; the lower signal-to-noise ratio in these data resulted in numerous 'spikes' in the IRIS interferograms, and a correspondingly high rejection rate during processing. Consequently, very few Martian spectra were obtained on odd numbered orbits. In addition, after 107 days (Lsubs:353), the changing planetary alignments moved the earth off of the transmitting antenna pattern; thereafter, only data that had been recorded were obtained, since transmission required a spacecraft maneuver to target the earth. Consequently, only a handful of data were received from Lsubs:5 to the end of mission at Lsubs:98.

Confidence Level Overview : Interferograms with serious distortions or uncorrectable spikes have been rejected during processing. These deletions, together with the calibration observations, result in gaps in the stream of Martian spectra.  The IRIS instrument measures net infrared flux, either emitting radiation (when viewing a target colder than the instrument), or absorbing radiation (when viewing a target warmer than the instrument). The sign (phase) of this flux is calculated as a function of wave number, and is applied during calibration to provide the proper amplitude of the spectral radiance. Errors in this phase determination sometimes occur at wavenumbers where the brightness temperature of the scene is very close to the instrument temperature (~254 K). This produces small discontinuous 'steps' in the resulting radiances. A few such errors remain in the dataset, usually in the brightness temperature range 252K - 256K.  The noise equivalent spectral radiance (NESR) provides a measure of the random errors in the spectra, expressed in radiometric units. It is defined as the radiance corresponding to a signal-to-noise ratio of unity, and represents the one-sigma uncertainty in an individual spectrum. It is calculated from the standard deviation of measurements taken while the instrument is viewing deep space. Above 250 cm**-1, the NESR is typically 5.E-8 Watts/cm**2/sr/cm**-1; actual values of the NESR are given for selected wavenumbers in the following Table 1 (above 800 cm-1, the corresponding Noise Equivalent Brightness Temperature increases roughly linearly, from approximately 140K at 1000 cm-1, to 240K at 2000 cm-1). A complete NESR spectrum is included in the auxiliary data of the MARINER9-MARS-IRIS-3-RDR-V1.0 dataset. Six sharp spikes, with amplitudes typically 10-50% of the NESR, are present in the NESR spectrum. Such spikes can represent the transforms of harmonic interference. The wavenumber of each spike, along with the interference frequency required to produce the spike, is shown in Table 2. The most probable sources of these spikes are transients caused by the engineering telemetry channels which have characteristic frequencies of 8-1/3 and 33-1/3 bps. The source of the interference at 28.27 Hz is unknown.   Table 1. IRIS NESR (in 10.E-8 Watts/cm**2/sr/cm**-1).  Wavenumber 200. 225. 250. 300. 350. 400. 800. 1200. 1600. 1958.  NESR 50. 10. 6. 3.5 5. 3. 5. 6. 5. 5.   Table 2. Interference spikes in the IRIS NESR.  Wavenumber Frequency Probable Source (cm**-1) (Hz)  356. 8.36 8-1/3 bps (telemetry rate) 713. 16.76 2*(8-1/3) 1069. 25.12 3*(8-1/3) 1203. 28.27 unknown 1426. 33.52 4*(8-1/3); 33-1/3 1782. 41.88 5*(8-1/3)   At wavenumbers where the radiance of the scene is near or below the instrument noise level, e.g., at high wavenumbers in polar spectra, the calibration procedure has resulted in nonnegative radiances. Consequently, averaging of spectra over such intervals does not result in a statistical reduction of noise, but tends toward a small, nearly constant, positive value. This is meaningless.  The pointing information provided in the headers was derived by the Mariner 9 Project from knowledge of the spacecraft position (determined from trajectory analysis), the spacecraft orientation (determined from sun and star sensor data, and by displacements within the limit cycle), and the articulation of the scan platform on which the instruments were mounted. Early in the mission, while the orbit was being adjusted frequently, such information was inadequate to derive the detailed pointing of the instruments, and so is missing. Gaps in the pointing information can be found in the headers for orbits 1 and 2. The orbit was stabilized with a periapsis near 1390 km by orbit 5 (it varied by a few kilometers thereafter, due to irregularities in the Martian gravitational field). At orbit 90(?), periapsis was raised to a value near 1650 km, where it remained until the end of the mission. Thus, spacecraft altitudes significantly below these values are spurious; other related geometric variables may also be corrupted.  Scattered throughout the dataset are 110 records for which the phase angle at the observation point (PHASE_ANGLE) is zero, and 7 records for which the spacecraft altitude (SC_ALT), angle between the line of sight and the center of Mars (ANGLE_LOS_MARS), angle between the line of sight and the nearest limb point of Mars (ANGLE_LOS_LIMB), and the angular semidiameter of Mars (ANGLE_SEMI_DIAM) are also all zero; other geometric variables for the observation are present. For all geometric computations, the radius of Mars is taken as 3387 km. The angular radius of the IRIS FOV is taken as 2.20 degrees.  At various times during the mission, some of the parameters in the header were given meaningless values; these values have been replaced with estimated values, and a flag has been set to indicate that a change has been made.