Explanatory Notes for Polarization Data
Summary data tables for the spectropolarimetry and imaging polarimetry
can be found from http://james.as.arizona.edu/~psmith/Fermi/DATA/data.html.
The following is a description of the data columns presented.
Links to the full-resolution linear polarization Stokes spectra can be
found in the summary tables for individual nights or from the links
to individual objects or from http://james.as.arizona.edu/~psmith/Fermi/DATA/specdata.html.
Column description for the polarimetry summary tables:
- (1) Object name (also see the target lists).
- (2) Date (in the Greenwich time zone)
of the observation in yyyy-mm-dd format.
- (3) Universal Time of the observation in hh:mm:ss format. This time refers to the
approximate midpoint of the observation.
- (4) Julian Day as calculated from columns (2) and (3).
- (5) Wavelength range of the listed polarization measurement. For
spectropolarimetry, this consists of a range given in Angstroms.
In the case of imaging polarimetry, the filter bandpass designation is given.
- (6) The median value of the q Stokes parameter measured from
the spectropolarimetry within
the wavelength range listed in column (5).
For imaging polarimetry, the measured q is listed.
- (7) The root mean square of the deviations of the q measurements
for the pixels within the wavelength range listed in column (5) for
spectropolarimetry. For imaging data, this column gives the uncertainty
in q based on photon statistics.
- (8) The median value of the u Stokes parameter measured from
the spectropolarimetry within
the wavelength range listed in column (5).
For imaging polarimetry, the measured u is listed.
- (9) The root mean square of the deviations of the u measurements
for the pixels within the wavelength range listed in column (5) for
spectropolarimetry. For imaging data, this column gives the uncertainty
in u based on photon statistics.
- (10) The fractional linear polarization measured in percent. Essentially,
P = 100 * SQRT[<q>2 + <u>2], but P has been corrected
for the statistical bias associated with the parameter being a positive,
definite quantity (see Wardle & Kronberg 1974). Generally, this correction is very
small given the high S/N of most of the data presented.
- (11) The uncertainty (in percent) in P based on the values given in columns (7) and (9).
- (12) The position angle on the sky of the linear polarization given
in degrees [θ = 0.5 * tan-1(<u>/<q>)],
with 0º (=180º) indicating that the polarization is aligned
north-south. The position angle is measured in the usual astronomical sense:
the angle increases going east of north.
Note that there is a 180º ambiguity in θ (e.g., 90º = 270º,which indicates, in this case, an east-west orientation of the polarization vector).
- (13) The estimated uncertainty in θ based on the measurement
of P (10) and its uncertainty. For a polarization measurement with P/σ(P) > 3, a good approximation for the uncertainty in θ is by,
σ(θ) = 28.65 * σ(P)/P (in degrees). This is the value
listed in column (13). For lower S/N, the estimate of the position angle
uncertainty becomes more complex and somewhat model dependent. However, at these
low S/N levels, it is probably best to simply ignore the listed value of
θ and consider the quantity as undefined.
Also, note that the quoted uncertainty does not include any systematic
error in the grid of polarimetric standard stars used to
calibrate θ. This may result in an offset of slightly more than a
degree between observing runs.
Changes in θ for an object within an observing run, however,
can be measured to
a much higher degree of accuracy, and for this purpose, we quote the uncertainty
based on photon statistics and the level of polarization observed.
- (14) A link to the full resolution, wavelength-calibrated
normalized q spectrum that can be used for
different polarization analyses, binning, etc. than presented in the
summary table. Use the "Save Link as" option to download the FITS file.
See the README file for the
spectroscopic observations for a detailed description of these data.
- (15) A link to the full resolution, wavelength-calibrated
normalized u spectrum that can be used for
different polarization analyses, binning, etc. than presented in the
summary table. Use the "Save Link as" option to download the FITS file.
See the README file for the
spectroscopic observations for a detailed description of these data.
Please note that no correction for interstellar polarization has been made to these data.
Slit width, exposure duration, airmass and epoch of the observation can
be found in the FITS headers of the Stokes spectra. The 'APERTURE' values correspond to the following
slit widths for the Bok and Kuiper telescopes: 1 = 2", 2 = 3", 3 = 4.1", 4 = 5.1", 5 = 7.6", 6 = 12.7".
Please contact Paul Smith (pssmith@arizona.edu) if you have questions concerning these web pages.