Cartesian Coordinates

For accurate work, Cartesian (XYZ) coordinates are much to be preferred, since the use of latitude, longitude and height entails a number of possible uncertainties and ambiguities. In particular, there are two families of commonly-used geodetic coordinate systems (lat/long/height) that can differ by as much as 200m. (The fact that the ATCA site survey in 1984 fell into just this trap emphasises the point.)

The choice of the correct type of geodetic coordinate depends on the application. In most cases, the geocentric WGS84 given below will probably be appropriate (eg in the calculation of LST). The non-geocentric system AGD84/AHD currently used by Australian cartographers and surveyors is unlikely to be relevant for astronomy, but is included for completeness and to allow comparison with existing physical surveys.

Ref	X(m)		e(X)	Y(m)		.	e(Y)	Z(m)		.	e(Z)
W196	-4750915	.	837	$\sim$2	2792906	.	182	$\sim$2	-3200483	.	747	$\sim$2
W196_vlbi	-4750924	.	9	$\sim$2	2792911	.	5	$\sim$2	-3200489	.	9	$\sim$2
MOPRA	-4682768	.	1	$\sim$1	2802618	.	8	$\sim$1	-3291759	.	2	$\sim$2
PARKES	-4554231	.	533	.031	2816759	.	109	.018	-3454036	.	323	.025

Notes: The coordinates are in a conventional geocentric right-handed Cartesian system, with the Z axis parallel to the mean rotation axis (North) and the X-axis aligned towards the intersection of the Greenwich meridian and the equator.

W196 refers to the centroid of the four pier-tops of station W196 of the ATCA. The listed position is that currently in use by CAOBS in the STATION.COORDINATES file. The position is derived from a physical survey (pre-GPS: see ATNF documents AT/15.6.1/001-003, AT/15.6.2/012) and is uncertain at the $\sim$2m level although the relative positions of the CA stations are accurate at the $\sim$mm level. The reference frame of the survey is unstated, but is probably AGD66. However, for consistency with existing coordinates in use at the ATCA it is assumed to be AGD84 ($\sim$2m differences against AGD66). The geoidal height used in the survey (N[ANS]=11.4m) has also been retained for consistency (though more accurate estimates are since available).

The antenna coordinates used by CAOBS for array pointing prior to October 1991 were all approximately 200m in error (mostly a uniform translation) owing to a misconception of references systems in the surveyors' report. The coordinates currently in use are consistent with the position given here for W196.

Station W196 is the logical choice to define a single position for the entire array. It is exactly midway between the two outer stations (W0 and W392) and is the only station for which the four support piers lie in the true local horizontal plane to make the antenna azimuth axis strictly vertical. It should therefore be used to define local sidereal time (LST) for the whole array so that all antenna azimuth and elevation angles are consistent. (The azimuth axes of antennas on the other stations are all parallel to that of station W196 and are therefore progressively tilted from the vertical.)

W196_vlbi refers to a point 10 metres above the centroid of the pier tops, approximating the position of the intersection of the elevation and azimuth axes (relevant for VLBI). A more accurate estimate of the geoidal height has also been used (N[ANS]=13.5m or N[WGS84]=29.7m, based on OSU86E). This position will be improved by further VLBI observations (once an H-maser is available).

The MOPRA and Parkes positions both derive from VLBI measurements and hence refer in both instances to the point of axial intersection. The Parkes position was obtained from S/X Mark-III VLBI data in January 1992 (Parkes-Hobart-DSS45) and is nominally on the ITRF frame. The Mopra position was derived from 4.8GHz Mark-II data from May 1993.