Import FITS-IDI EOP information #1307
Comments
|
After privvate discussion with @walterfb we concluded that having the information about whether the EOP values are predicted, intermediate or final available in the table is indeed desirable. My proposal is to use the same terminology that is used in the IERS bulletins for this, so "predicted" (bulletin A), "preliminary" (bulletin B) or "final" (bulletin B). EARTH_ORIENTATION
TIMETimestamp of the EOP values; this is expected to be the UTC time although some sources use TAI instead UT1_UTCDifference between UT1 and UTC (UT1 - UTC) in seconds. PMPosition of celestial pole; these are provided as X and Y components (in that order) Rationale: UT1_UTC and PM_x and PM_y are the parameters used in the IAU 2000 and 2006 precession and nutation models. Older models used more parameters. These models are no longer in widespread use in VLBI and we don't plan to implement support for these models in CASA. TYPEType of the EOP value; one of the strings PREDICTED, PRELIMINARY or FINAL or an empty string if the type is not known. |
|
When implementing concatenation of MSes that have an EARTH_ORIENTATION table, I realized that these MSes could have EOP values for the same time stamps that are slightly different. So we need an OBSERVATION_ID column to be able to distinguish between them in the merged MS. EARTH_ORIENTATION
TIMETimestamp of the EOP values; this is expected to be the UTC time although some sources use TAI instead. OBSERVATION_IDObservation idientifier. UT1_UTCDifference between UT1 and UTC (UT1 - UTC) in seconds. PMPosition of celestial pole; these are provided as X and Y components (in that order). Rationale: UT1_UTC and PM_x and PM_y are the parameters used in the IAU 2000 and 2006 precession and nutation models. Older models used more parameters. These models are no longer in widespread use in VLBI and we don't plan to implement support for these models in CASA. TYPEType of the EOP value; one of the strings PREDICTED, PRELIMINARY or FINAL or an empty string if the type is not known. |
Import Earth Orientation Parameters (EOPs) from FITS-IDI files and convert them into an EAUTH_ORIENTATION subtable as documented in issue casacore#1307. These EOPs are necessary for applying corrections in case more accurate values become available after correlation. This is import for astrometric VLBI observations.
Import Earth Orientation Parameters (EOPs) from FITS-IDI files and convert them into an EAUTH_ORIENTATION subtable as documented in issue #1307. These EOPs are necessary for applying corrections in case more accurate values become available after correlation. This is import for astrometric VLBI observations.
Astrometric VLBI requires a very accurate correlator delay model. The calculation of the delays involve a terrestial-to-celestial transformation which depends on a number of parameters that describe the orientaton of the earth: the difference between UT1 and UTC, and the position of the celestial pole. These Earth Orientation Parameters (EOPs) can be predicted, but ultimately need to be measured for maximum accuracy. The most accurate measurements are only available after a few weeks and can differ significantly (for astrometric purposes) from the predictions. These final measurements may not be available at time of correlation, so data may be correlated with predicted or less accurate measurements. As long as these difference with the final values is small, phase corrections can be made after correlation. Code to apply these corrections is being developed for CASA.
To calculate these corrections we need to know the EOPs that were used for correlation. For the VLBA (and potentially other VLBI arrays that use the DiFX correlator) these values are available in the (nonstandard) CALC table in the FITS-IDI files. So we need to import these values and make them available in the MeasurementSet. A time series of EOPs is provided, typically extending two days before and after the observation such that accurate interpolation (in time) of the values is possible. This issue proposes a new MeasurementSet extension table (in the spirit of the GAIN_CURVE and PHASE_CAL tables that have been added somewhat recently). Feedback on the proposed table is welcome.
I will be implementing support for importing EOPs from FITS-IDI files in casacore based on the outcome if that discussion.
EARTH_ORIENTATION
TIME
Timestamp of the EOP values; this is expected to be the UTC time although some sources use TAI instead
UT1_UTC
Difference between UT1 and UTC (UT1 - UTC) in seconds.
PM
Position of celestial pole; these are provided as X and Y components (in that order)
Rationale: UT1_UTC and PM_x and PM_y are the parameters used in the IAU 2000 and 2006 precession and nutation models. Older models used more parameters. These models are no longer in widespread use in VLBI and we don't plan to implement support for these models in CASA.
Further comments
It may be relevant to know whether the EOP values are predicted, intermediate or final. This could be done in the form of a string with reserved keywords like what is done for mount types in the MS 2.0 standard.
The text was updated successfully, but these errors were encountered: