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

    The Electron Reflectometer Data Record (ERDR) is a time ordered
    series of electron measurements from the Mars Global Surveyor (MGS)
    Mission. Each record consists of a time tag with 19 scalar data
    points representing measurements of the electron flux in 19
    different energy channels, ranging from 10 eV to 20 keV, with an
    energy resolution of 25%.  Each data point is a measure of the
    electron flux (cm-2 sec-1 ster-1 eV-1) averaged over a 360 x 14
    degree disk-shaped field of view (FOV).  During the Mapping Phase,
    as the spacecraft orbits the planet the ER field of view sweeps out
    the entire sky (4-pi ster) every 58.5 minutes, which is much longer
    than the integration time per record (2 to 48 sec, depending on
    energy and telemetry rate) and much longer than most timescales of
    interest in Mars' plasma environment.

    The ERDR is intended to be used in conjunction with MGS Magnetometer
    (MAG) data records, which provide the magnetic field vector and
    spacecraft ephemeris data as a function of time.  Electrons travel
    along the magnetic field lines in tight helices (few km radius) at
    high speed (roughly one Mars diameter per second).  Thus the
    electron data contain information about the plasma environment as
    well as the large-scale configuration of the magnetic field, which
    is sampled locally by the MAG.


  ====================================================================
  Parameters
  ==========

    The Mars Global Surveyor ER data set consists of a time ordered
    series of electron flux measurements in 19 energy channels, ranging
    from 10 eV to 20 keV.  The ER data are organized into ''packets,''
    each of which contains 12, 24, or 48 seconds of data, respectively,
    for high, medium, and low spacecraft telemetry rates.  Each packet
    is further subdivided into samples.  There are from 1 to 6 samples
    per packet, depending on the energy channel, as given in the table
    below.  The ER data set is generated at the rate of 6 samples per
    packet, regardless of energy.  When there are fewer than 6 samples
    per packet at a particular energy, data values are repeated in order
    to maintain a uniform table.  The time listed for each record is the
    center of the sampling interval.  When records are repeated, taking
    an average of the times for all repeated records provides the center
    time for that sample.

    The energy channels and sampling intervals are as follows:


        Channel Number        Energy Range     Samples per Packet
    ---------------------------------------------------------------
               0               13 - 20   keV           1
               1              8.0 - 12   keV           1
               2              4.9 - 7.5  keV           1
               3              2.9 - 4.6  keV           3
               4              1.8 - 2.8  keV           3
               5              1.1 - 1.7  keV           3
               6              680 - 1046  eV           6
               7              415 - 639   eV           6
               8              253 - 390   eV           6
               9              153 - 237   eV           6
              10               92 - 144   eV           6
              11               72 - 87    eV           2
              12               56 - 67    eV           2
              13               43 - 52    eV           1
              14               33 - 40    eV           2
              15               25 - 30    eV           1
              16               18 - 23    eV           2
              17               14 - 17    eV           1
              18               10 - 13    eV           2
    ---------------------------------------------------------------


    The ER has a 360 x 14 degree disk-shaped field of view.  The 360
    degrees are divided into 16 angular sectors, each with a separate
    counter that is read out in telemetry.  The sizes and look
    directions of these sectors are programmable to within an accuracy
    of 1.4 degrees.  For these data, the FOV is divided into 16 equally
    sized 22.5 x 14 degree sectors that remained fixed in spacecraft
    coordinates.

    Throughout pre-mapping, some of these sectors were masked because of
    obstructions in the FOV, most notably the stowed high gain antenna
    (HGA), which blocked 3 sectors.  Three additional sectors are
    partially obstructed by the -Y solar array gimbal/yoke assembly and
    corners of the spacecraft bus.  Apart from the HGA, these
    obstructions are minimal, so data from these sectors are still
    considered to be of good quality.  After HGA deployment just prior
    to mapping, the entire FOV became useable.  An in-flight calibration
    was then performed to determine the relative instrumental
    sensitivity around the FOV to an accuracy of about 10%.  For this
    mapping data set, we average data from all 16 sectors to form a
    scalar ''omnidirectional'' value.


  ====================================================================
  Processing
  ==========

    Processing is carried out at the Space Sciences Laboratory (SSL) of
    the University of California, Berkeley, (UCB) to convert the raw
    data to measurements of the omnidirectional electron flux (cm-2 s-1
    ster-1 eV-1).  Because of the instrument's high dynamic range (six
    decades), the onboard digital processing unit (DPU) compresses the
    raw counts in a logarithmic scale.  The first step is to decompress
    the raw counts and construct a three-dimensional data array, where
    the first dimension is time (6 elements per telemetry packet), the
    second dimension is direction around the FOV (16 elements), and the
    third dimension is energy (19 elements).

    The next step is to sum over the 16 angular sectors to produce a
    two-dimensional time/energy array.  Raw count rate (R) is then
    obtained by dividing the raw counts by the integration time (0.0625
    sec per energy step).  The data are next corrected for deadtime.
    During the time it takes the instrument to process a single electron
    (known as the ''deadtime'', which is about 0.4 microsec for the ER),
    it ignores any other electrons.  The raw count rate is multiplied by
    the factor 1/(1 - RT), where T is the deadtime, to obtain corrected
    count rate.  Data values are masked (set to -9.999e-9) when the
    deadtime correction factor exceeds 1.25.  These data are NOT
    CORRECTED for a background count rate due to cosmic rays and noise
    in the electronics (about 10 counts/sec).  Most of the time, the
    signal in the highest energy channel (13-20 keV) is dominated by
    background.  Exceptions to this sometimes occur during bowshock
    crossings or during energetic solar events.  Assuming that the
    highest energy channel contains 100% background, the background
    level for the lower energy channels can be estimated as follows:

      Channels  0-10 (92 eV - 20 keV): B(E) = B(20 keV) * (20 keV/E)
      Channels 11-18 (10 eV - 87 eV):  B(E) = B(20 keV) * (20 keV/E) *
        43.5

    where B(E) is the background level (in units of cm-2 s-1 ster-1
    eV-1) at energy E.  The background is typically negligible at
    energies below about 1 keV.  Background correction is essential at
    higher energies.

    Data are collected through two separate apertures that differ in
    their transmission by a factor of 43.5.  At low energies (10 eV to
    87 eV), the smaller aperture is used to attenuate high fluxes, and
    at high energies (92 eV to 20 keV), the larger aperture is used to
    maximize the sensitivity to low fluxes.  The corrected count rates
    in energy channels 11-18 (10-87 eV) are multiplied by the factor
    43.5 to compensate for the smaller aperture size.  Finally, we
    divide by the geometric factor (0.02 cm2 ster) and the center energy
    (eV) to obtain the differential flux (cm-2 s-1 ster-1 eV-1).

    The data are organized into a table with a uniform time step for all
    energy channels.  Since the sampling interval is different for
    different energy channels, data values are repeated within each
    packet, as necessary, to enforce a uniform time step.  The data are
    sent via FTP to the Principal Investigator (Mario Acuna) at Goddard
    Space Flight Center (GSFC), where they are incorporated with the
    magnetometer data.


  ====================================================================
  Data
  ====

    The ERDR data set consists of a single time-ordered table.  Each
    record contains a time stamp and 19 data values, representing the
    omnidirectional electron flux in 19 different energy channels
    ranging from 10 eV to 20 keV.


  ====================================================================
  Ancillary Data
  ==============

    No additional ancillary data is required beyond that described for
    the MAG.


  ====================================================================
  Coordinate System
  =================

    The data are presented in omnidirectional format.  The time tags
    contained in the ER data set should be used to obtain the
    corresponding spacecraft ephemeris information from the MAG data
    set.


  ====================================================================
  Software
  ========

    Data reduction software for the ER is written in IDL.


  ====================================================================
  Media/Format
  ============

    The ER data are provided in the form of 20-column ascii tables.
    Storage media are described in the MAG documentation.

====================================================================
  Confidence Level Overview
  =========================

    The ER is mounted on the spacecraft body, where measurements are
    susceptible to spacecraft charging and FOV blockage.  This ER
    instrument design is typically used on a rapidly spinning spacecraft
    (few seconds period), on which the disk-shaped FOV would sweep out
    the entire sky in a time that is short compared with most
    timescales of interest.  However, since MGS spins slowly (once per
    orbit), each data record covers only a small region of the sky.
    Despite this limitation, the scalar flux provided in the ERDR is
    suitable for identification of plasma boundaries (bow shock,
    magnetic pile-up boundary, ionopause) and following the evolution of
    the electron energy distribution, which is useful for evaluation of
    the plasma environment and interpretation of the magnetometer data.
    Any application of these data that requires an unbiased average over
    all look directions (4-pi ster) is NOT RECOMMENDED.


  ====================================================================
  Review
  ======

    The ERDR will be reviewed internally by the MGS MAG/ER team prior to
    release to the planetary community. The ERDR will also be reviewed
    by PDS.


  ====================================================================
  Data Coverage and Quality
  =========================

    ER data are recorded continuously.  Data coverage depends almost
    entirely on the fraction of the spacecraft telemetry that can be
    received by the DSN.  The mapping orbit lies close to the ionopause
    altitude.  Because of spatial and temporal variations in the
    ionopause, the ER can sample several different plasma environments,
    including the ionosphere, the magnetosheath, the magnetotail, and
    closed magnetic field lines anchored to remanent crustal sources.
    Data quality is best when the spacecraft is within the planet's
    shadow.  In sunlight, data quality is a function of spacecraft
    rotation phase, since photoelectron contamination depends on the
    illumination pattern.


  ====================================================================
  Limitations
  ===========

    The ER is mounted on the spacecraft instrument deck and has a
    disk-shaped FOV that is orthogonal to the spacecraft XY plane and
    nearly orthogonal to the spacecraft Y axis.  (There is a 10-degree
    rotation about the Z axis to minimize spacecraft obstructions in the
    FOV.)  This 360-degree FOV is divided into 16 angular sectors, each
    22.5 degrees wide.  Throughout mapping, the ER is in
    ''fixed-sector'' mode, meaning that these 16 angular sectors
    remained constant in the spacecraft reference frame, sweeping out
    the entire sky every 1/2 of an orbit.

    Parts of the spacecraft are within the instrument's FOV.  The high
    gain antenna (HGA), which blocked ~70 degrees of the FOV during
    aerobraking is not in the FOV during mapping.  Smaller amounts of
    blockage are caused by attitude control thrusters and the -Y solar
    array gimbal and yoke assembly.  One effect this has on the
    measurements is to block ambient electrons from the directions of
    the obstacles. This is most clearly seen at high energies (> 100
    eV), which are only slightly deflected by the spacecraft floating
    potential.  In addition, when these obstacles are illuminated by the
    sun, they emit photoelectrons up to ~50 eV, which can enter the ER
    aperture and elevate the counting rate at low energies.  The
    detailed signature of this effect depends on the illumination
    pattern as the spacecraft rotates, which is a function of the angles
    between Earth, Mars, and the Sun.  These angles varied throughout
    the mapping phase. Photoelectron contamination has not been removed
    from the data; however, the presence of contamination is readily
    identified in the low energy channels (< 50 eV) by a sharp (nearly
    discontinuous) increase in counting rate which appears at regular
    100-minute intervals.  The contamination disappears as abruptly as
    it appears.

    For a duration of ~4 minutes every 1/2 spacecraft spin (when the
    spacecraft illuminated) sunlight can directly enter the ER aperture
    and scatter inside the instrument, creating secondary electrons.
    (Note: the spacecraft spins once per orbit to keep the nadir deck
    pointed at the planet.)  A tiny fraction of these photons and
    secondary electrons can scatter down to the anode and create a
    ''pulse'' of spurious counts.  This sunlight pulse appears at all
    energies, but is most noticeable from 10 to 80 eV and above 1 keV.
    Sunlight pulses have not been removed from the data.

    The instrument's energy scale is referenced to spacecraft ground.
    In sunlight, spacecraft ground floats a few volts positive relative
    to the plasma in which the spacecraft is immersed.  Electrons are
    accelerated by the spacecraft potential before they can enter the ER
    aperture, thus all energies are shifted upward by a few eV.  In
    addition to shifting the electron energy, the trajectories of low
    energy electrons can be significantly bent by electric fields around
    the spacecraft.  Thus, the energy scale and imaging characteristics
    are relatively poor at the lowest energies (10-30 eV), becoming much
    more accurate at higher energies.