Finding your Position

November 27, 2017

by Peter Anderson

When you set up your Equatorial telescopes you must know your latitude and (for Southern hemisphere observers) the direction of true south. For certain astronomical observations (Lunar and Asteroidal Occultations etc), it is essential to know the longitude and latitude to considerable accuracy. This is essential for observers planning to observe grazing lunar occultations and total solar eclipses.

Older Australian maps prior to 1966 relied on different datums. I recall in Brisbane that the ‘Sydney Datum’ was apparently referenced. When the Australian Geodetic Datum was introduced in 1966 a 'block shift' was needed to be applied to positions. For South East Queensland this was accomplished by subtracting 0.2" (0.2 of an arc second), from the scaled latitude and 5.25" from the scaled longitude. After this adjustment the position shown on the maps was accurate to 0.2". In effect this moved the position of my station in Brisbane some 144 metres just slightly north of west.

In 1966 the Australian National Spheroid was adopted and this was the mathematical model representing the part of the Earth covering Australia. Much hard preparatory surveying work had been done and the adoption of the new resulting datum was announced with considerable fanfare. This model was represented on the ground by a network of survey marks (brass plaques) with the origin being the Johnston Origin marker, the central geomagnetic station located in central Australia. It was optimized for Australia, of limited accuracy due to surveying and computing technologies at the time. It was known as the 1966 Australian Geodetic Datum (1966AGD).

In 1984, many of the original survey observations had been re-observed using laser technology and integrated with satellite technologies such as Doppler. In addition the Geoid was determined for the first time. The Geoid is the real-world shape of the Earth, ignoring topography, and based on sea level. This datum used the Australian National Spheroid as it did in 1966. It merely had a more accurate representation on the ground and the adjustment for my observatory between the 1978 South East Queensland determination when my site was surveyed and the 1984 update amounted to a movement of a mere two metres to the south south west (still within my observatory building). Understanding the geoid (in particular the Deflections of the Vertical) provided the first opportunity in 1984 to compute astronomical positions from coordinates derived from the geodetic networks.

On 1 January 2000, the Geocentric Datum of Australia (GDA94) came into effect. This is a totally new datum based on the centre of the Earth, optimized for a global solution and taking full advantage of the latest in satellite positioning and computing technologies. It is based on the GRS80 ellipsoid and is represented by the Australian Fiducial Network (AFN), (comprising eight permanent, continuously operating GPS receivers on the Australian mainland and Tasmania) in conjunction with additional sites beyond the Australian mainland.  Its precision is accurate to a few centimetres.  

For our astronomical positioning purposes the GDA94 positions are effectively the same as the World Geodetic System WGS84 currently used for the GPS satellite navigation network at the epoch of 1994.0.

The co-ordinates for this datum for my site (Taylor Range Observatory), represent a shift of some 200 metres to the north north-east at a bearing of 30 degrees from AGD66.

However, this is not the final word.  As indicated by its title, GDA94, this new datum is based on the position of Australia at 1 January 1994. The whole Australian part of the Tectonic Plate is gliding over the surface of the Earth in a NNE direction at a rate of around 7cm per year. GDA94 is a plate-fixed datum; that is the AFN coordinates define the datum at that point in time.  The WGS84 datum is fixed to the centre of the whole Earth and is therefore called an Earth fixed datum. Over time, the coordinates of a single point on the surface of the Earth, if described in WGS84, will progressively depart from those of the  apparently fixed GDA94 coordinates by this 7cm/ year. By 2020 the sum of this incremental creep will be around 1.8m!

Increasing accuracy is required for the use of the new generation of GPS receivers – think of driving and precise agricultural uses.  Therefore the movement between 1994 and 2020 of the tectonic plate required a further adjustment to the datum.  A 1.8 metre adjustment was made by Geoscience Australia on 1st January 2017 and is part of the ongoing changeover to the new official reference frame, called the Geocentric Datum of Australia 2020, (GDA 2020) replacing GDA 94.  This datum like the WGS84 is identical for our position determination purposes at epoch 2020.0 with the positions in GPS units and GPS satellite imagery (such as Google Earth).  There are two aspects to GDA 2020, one being the fixed datum reference as at 2020 (which is the position Australia will have moved to with the 1.8metre adjustment). The second aspect (which we would reference), is a moving one which henceforth keeps pace with the Australian specific tectonic plate movement to the north north east, and so remains totally consistent with the identical moving positions obtained from WGS 84 and the GPS and other satellite data. This second aspect is called the Australian Terrestrial Reference Frame (ATRF) and can be thought of as a regional realisation of the International Terrestrial Reference Frame (ITRF).

Next we come to the question of altitude, and unfortunately for technical reasons there are some variations:

The figure given for the altitude of my observatory averages 177 metres for GPS units.  I understand the reason is that these heights are only accurate to 5 to 10 metres without multipath.  

The altitude  of 175 metres was provided by the manual survey and shown on various topographic maps. I believe this figure to be totally accurate.

Google Earth places the altitude at 171 metres.  A quick mouse over of a Google Earth screen will be enough to quickly reveal altitude disparities of up to 5 metres and so this can be regarded as a close approximation.  However if I am required to provide a ‘Google Earth’ position, I will include their altitude figure for the sake of consistency.

In the past when engaging in grazing lunar occultation expeditions, a survey map of the target area where the graze was to occur was purchased. The predicted graze track was pencilled on this map.  The purpose was to establish temporary observing stations straddling the graze track and thus obtain a very accurate profile of the limb of the Moon. This was as accurate as the separation of the distances of these stations from each other relative to the track path.  Then the observers set out for the pre-arranged observing stations, but often had to compromise because of local difficulties.  After the event the actual observing sites were plotted on the map and their geographic positions carefully measured. Sometimes a station along a straight featureless street might have been chosen and then its position determined (say) by counting the power poles to an identifiable point. Much effort went into this. Nowadays handheld GPS units will provide sufficiently accurate positions instantly to enable quick plotting relative to the calculated path.

But it goes further than this.  For many years I quoted the position determined by the manual survey of my observatory site, and referenced the relevant datum. About a decade ago, the organisation to which I forward my observations queried this and advised that they now used ‘Google Earth’.   There was a difference of a handful of arc seconds between my figures and that of the Datum used by ‘Google Earth’.  As requested I immediately switched to ‘Google Earth’ referencing its datum WGS84.

Purists will point out that this datum WGS84 is Earth fixed and so the co-ordinates change by 7cm per year. Then there is a question of the accuracy of source images that Google uses to update from time to time. For our purposes an inaccuracy in their position of a metre or two and somewhat more in altitude is dramatically smaller than our best observational resolution to the extent that it would have insignificant effect on results. However, using Google Earth positions does provide an easily accessible standard reference, and if necessary any corrections can be applied later.

So determining a position to observe a ‘graze’ could simply become a matter of reading it out from a ‘mouse over’ on ‘Google Earth’, and with the added benefit of ‘Google Street View’ graze observers might survey a suitable site remotely – with enough space to set up and free from overhanging trees etc. They could then simply arrive with their equipment without need for any further locational support.

In summary, life is much easier now...

I wish to express my appreciation for his invaluable assistance to Dr Craig Roberts, Senior Lecturer, Surveying and Geospatial Engineering, School of Civil & Environmental Engineering, University of New South Wales, SYDNEY, AUSTRALIA

For further information and frequently asked questions, please refer:

There are related issues for observational astronomers, namely time determination and observational and instrumental and techniques, but each of these is a subject by itself.


AGD 1966 Johnston marker

Back to News List