PDS_VERSION_ID     = PDS3

RECORD_TYPE        = STREAM
OBJECT             = TEXT
  NOTE             = "Known errors and/or anomalies in the volumes"
  PUBLICATION_DATE = 2010-12-22
END_OBJECT         = TEXT
END

ERRORS AND/OR ANOMALIES IN THE CURRENT VOLUME

Volume CORADR_0214, Version 02
------------------------------

1.  The Cassini Radar Transition file (EXTRAS/CRT_214_V02.TAB) contains
no information about ScanStart and ScanEnd transitions.

2.  The uncompressed LBDR and BIDR products have attached labels.
Normally, a file that has been compressed with ZIP would have been
generated without an attached label.

3.  In the volume index table (INDEX/INDEX.TAB), double quotes enclose
all the date/time values.  Normally, PDS date/time values are not
quoted, but quoting makes parsing easier for some applications.

4.  The HTML documents in the DOCUMENT directory contain HTML character
codes that, while all-ASCII, are not easily interpretable by someone
who is reading the HTML documents as text documents.  For example,
"α" represents the lower-case Greek character "alpha" and is
rendered as such in a web browser.  Equivalent and more legible character
codes (e.g., "&#alpha;") are available as of the HTML 4.0 specification
but cannot be used here, as PDS requires HTML documents to comply to the
HTML 3.2 specification.

5.  Antenna temperature, brightness temperature, and receiver temperature are
defined in Janssen, M. A., "An Introduction to the Passive Microwave Remote
Sensing of Atmospheres," Chapter 1 in Atmospheric Remote Sensing by
Microwave Radiometry, (M. Janssen, ed.), pp. 1-35, Wiley & Sons,
New York (1993).  The archived value in the SBDR and LBDR files gives
uncalibrated antenna temperature in units of Kelvin.
The best current algorithm for correcting the archived antenna temperatures is
Ta_corrected = Ta_archive * ( 0.920 - 0.0041*( t - 1.90 ) )
where t = time in years and fractional years since 2004.0
(0 UTC on 1 Jan 2004)
This algorithm is based on the radioastronomical flux scale at 2-cm wavelength
by direct comparison of distant (unresolved) Titan measurements by the Cassini
radiometer with VLA measurements of Titan reported by Butler and Gurwell,
2004.  This algorithm will improve with time as more distant Titan
measurements are obtained and as more radioastronomical comparison sources are
included.
Butler, B. J., Gurwell, M. A. 2004. Radio Wavelength Observations of Titan
with the  VLA. Bull. Am. Astron. Soc. 36, 6.04.
This algorithm applies also to all preceeding volumes.

6. Ideally the calibrated antenna temperature is referenced to cold sky at
2.7 K, although no guarantee is made that this zero-level accounting has been
correctly made.  Also, the antenna temperature is defined for this application
as just the average brightness temperature in the measured beam out to 2
degrees from the beam axis, and does not allow for possible contributions from
the far sidelobes (sidelobes outside of 2 degreees).  The archived value will
include an additional contribution if the  far sidelobes happen to fall on 
other than cold sky.   In particular, there is an offset to be expected and 
accounted for when an extended source like Titan or Saturn is observed from a 
close distance.  A more detailed explanation will be found in Janssen et al.,
2009. System gain is the quantity that multiplies the raw sky counts to 
convert to the uncalibrated Kelvin scale.
Receiver_temp is the receiver noise temperature Tr (comparison made at
internal reference switch).   The receiver temperature plus the antenna
temperature is equal to the total signal (raw counts times system gain).

7. ant_temp_std is a measure of the rms uncertainty of Ta, and is only an
estimation.  It is obtained as the standard deviation of Ta for three points,
Ta(I-1), Ta(I), and Ta(I+1), and is a useful measure that identifies
questionable data.  For example, it gets large when the beam is sweeping
across a brightness discontinuity.

8. All BIDRs except for the byte-valued backscatter images (BIB*.IMG) were
produced by JPL. The byte-valued backscatter images were produced by USGS.
The USGS BIDRs have unit strings in their labels which are all CAPS. The 
other BIDRs have lower case units. According to the PDS dictionary, both
are acceptable.

9. Checksums were not computed for floating point valued BIDRs. The 
CHECKSUM keyword was assigned a zero value.

10. In the primary BIDR images a technique was used to remove systematic 
errors due to thermal noise and BAQ compression during downlink. The
technique appears to have a slight negative bias for very low SNR, i.e.
when the signal power is less than one tenth the thermal noise power. 
Without noise subtraction, backscatter values are artificially high and 
always positive. With noise subtraction, negative values occur both due 
to the aforementioned bias and due to residual random error. Byte-valued 
BIDRs are produced by transforming all data to Decibels by taking the 
logarithm and then  multiplying by 10. Data is clipped below a minimum 
value (typically -20 dB) determined by the OFFSET  keyword in the 
attached label. All negative backscatter values or value belows 0.01 
(-20 dB)are assigned to the minimum pixel value. Anyone interested in 
viewing data below this threshold needs to use the REAL-VALUED 
backplanes. 

  
11. Typically SAR observations are only obtained for Titan. The
Rhea observation contained in this volume is one of only a
handful of non-Titan Cassini SAR observations. As such, there are
some differences in the data from nominal Titan SAR observations. 
The remainder of the list describes those differences.

12. In order to estimate the normalized backscatter cross-section from
the returned echo power in each SAR pixel it is necessary to divide by the
integral of the two-way antenna gain divided by the range to the fourth power.
This integral is performed over the area on the surface of Rhea for 
each SAR pixel. For nominal Titan SAR, pixels are small compared to the gain 
pattern and rectangularly-shaped so that a simple 
approximation can be used without resorting to a full numerical integration.
For Rhea the full integration was performed for each pixel with a 500 m
step size.

13. Geolocation on the surface of Rhea makes use of the nominal
triaxial shape of Rhea (radx,rady,radz)=(767.2 762.5 763.1) km rather 
than assuming a sphere as was done for Titan. Nonethless in order to simplify
the map projection the images were projected onto a 764.2695 km radius sphere
as described in the labels for each BIDR file. These coordinates are not
the same as the planetodetic coordinates reported in the LBDR and SBDR files.

14. The correction used to remove incidence angle variation in the BIF*.IMG
and BIB*.IMG files was different than that used for Titan. Each pixel was
multiplied by a function of incidence angle f(I)=1.6930/(2.15*cos(I)^1.45)

15. For all observations archived after Dec 1, 2010, the geolocation code was
modified to account for the effect of special relativity on Doppler shift. 
Without this change, the geolocation would have been in error by 1.5 pixels 
for Enceladus SAR. This error corresponds to 3 parts in one hundred thousand 
of the overall Doppler shift of the echo. Such a high level of fidelity was 
not needed for previous volumes in which the geo-location error due to 
relativity was much less than a pixel width. Now that it has been implemented,
we employ this correction for all observations even when it is too small to 
be significant.


16. The act_incidence_angle parameter in the SBDR and LBDR was redefined to
be consistent with the incidence angle used to compute the incidence angle 
corrected sigma0. The boresight incidence angle value normally reported does 
not represent the wide range of incidence angles observed within the antenna
footprint. Instead, an effective incidence angle is reported that is a 
weighted average of the incidence angles observed. It is weighted by the 
estimated distribution of the return energy over incidence angle.

17. Pixels with SAR ambiguity contributions greater than 30% are marked as 
invalid. This includes pixels with contributions from so-called mirror 
ambiguities that share the same Doppler and range to target as the primary 
imaging area but come from opposite sides of the spacecraft velocity vector. 
Note in particular the invalid strip of data which bisects the image. This
omission is similar to what has been done in the past for Titan observations,
but differs from what was done for the single previously occurring Enceladus
SAR observation.

 

ERRORS AND/OR ANOMALIES IN PREVIOUS VOLUMES

Not Applicable