Australian Geodetic Datum 1966 : Fifty Years On
The Australian Geodetic Datum 1966 (AGD66), was adopted by the National Mapping Council on 21 April 1966, and proclaimed later that same year in the Commonwealth Gazette No. 84 of 6 October.
As far back as 1901 the learned professions started calling for the topographical mapping of Australia to be commenced. From the Conference of Surveyors General in Melbourne during May 1912 proposals emerged for a National Geodetic Survey of Australia. This conference recognised that such a geodetic survey was absolutely necessary for the production of accurate maps. Figure 1 shows the status of State triangulation as reported at the 1912 Conference. This work however, varied in its quality and was generally based on local datums. Until the early 1960s, the origin for most of the Australian Geodetic Survey was the coordinates of the Sydney Observatory. At the time of computation of the new national datum, various sections of its underlying survey were based on 20 different geodetic datums using four different figures for the earth. For example, the Western Australian survey was based on Perth Observatory, the surveys in northern Queensland on local origins at Cairns and Townsville while the Stuart Highway Traverse in the Northern Territory was computed in terms of an origin at Tennant Creek.
Figure 1 : Status of State triangulation as reported at the 1912 Conference.
Despite continued efforts to focus the governments of the time on this national need for a uniform reference system, little happened as is evidenced by this article by a staff Correspondent in Canberra for the Sydney Morning Herald. Published in the edition of Friday 24 January 1947 on page 2, under the banner Plan To Map Australia With Help of Radar, the article concluded an interesting fact which shows the need for accurate mapping of the Australian continent is that, as the result of the simultaneous exchange of relevant data between Mount Stromlo and Greenwich observatories, it has been found that some of the longitudes in Australia, previously regarded as accurate, are incorrect by approximately one-tenth of a second of longitude about 100ft, which is of some consequence in the preparation of highly accurate maps.
Australian Geodetic Datum 1966 (AGD66)
The geodetic survey of Australia was finally completed during the approximate 20 year period 1945-65 with the majority of the work in the last ten years. Ford (1979) documented this effort which can be found via this link.
The 1966 national adjustment of the Australian Geodetic Survey comprised some 2,506 stations, 53,000 kilometres of Tellurometer traverse, the geodetic triangulation nets, and 533 Laplace stations. The Australian Geodetic Survey was particularly strong in azimuth due to the great number of Laplace stations along the traverse network. Refer Figures 2 and 3.
Figure 2 : Triangulation nets (hatched) and loops (blue trigs by National Mapping and green trigs by Royal Australian Army Survey Corps and States) of the Geodetic Survey as at 1966.
Figure 3 : Laplace stations of the Geodetic Survey as at 1966.
The methods and instruments used on the geodetic survey suggested that the standard error between adjacent stations of the originally adjusted net was between 3 to 4 parts per million both linearly and laterally. Closures of the 58 loops of the survey, indicated that for distances of about 1500 kilometres, the average misclosure had been reduced to about 2 parts per million or 3 metres. The average distance between stations contained in the basic adjustment was 30 kilometres and 3 to 4 parts per million of such distances is equivalent to about 10 centimetres, or about the size of your multimedia remote control.
Coordinates on the Australian Geodetic Datum presented a firm foundation on which lower order surveys and all mapping could be based. Furthermore, these coordinates provided a basis for a point reference system, on which any location in Australia, for example an oil well, a mineral deposit, a gravity station, or a light house, could be described in precise and unambiguous terms. For mapping, the Australian Map Grid (AMG) was introduced. The latitudes and longitudes of features was then represented by a system of rectangular grid coordinates of northings and eastings.
Before 1966 and up to the late 1970s, mapping in Australia was based on many heterogeneous datums including astro-fixes, not related to any datum. An analysis of sixteen 1:250,000 scale map sheets of the R502 series was undertaken to determine the datum difference, AGD66 to astro-fix. The sixteen sheets were well distributed over Australia, and showed differences in latitude ranging from -6" to + 6" and in longitude from + 9" to - 8". These values can be converted to metres using the approximation of 1" equaling 30 metres. This analysis supported the decision to base the 1:250,000 scale mapping in areas without geodetic control on astro-fixes. It had been anticipated that astro-fix control would be nearer to the finally accepted geodetic datum than to the existing datums adopted for isolated geodetic or lower order triangulation networks. Such datums might have contained large, and at the time unknown, deflections of the vertical at their origins. However, the use of astro-fix control was also based on their flexibility to meet operational mapping priorities.
This analysis meant that the R502 series of maps contained information that was correctly located to less than 1 millimetre at map scale relative to AGD66. In addition, by overprinting (in cyan) about half of the R502 sheets with the appropriate AMG metric grid, their information was immediately standardised (almost) to AGD66.
Although it is now fifty years since AGD66 was established, the AGD66 and a few follow-ons were replaced by GDA94; but even that datum is now over 20 years old. As Australia is part of the fastest moving continental tectonic plate on Earth, GDA94 coordinates are over 1.5 metres different to the International Terrestrial Reference Frame (ITRF) coordinates used by Global Navigation Satellite Systems, such as GPS. By 2020 this difference will approach 2 metres and will start to be noticeable, if not already, by sophisticated users.
The new 2020 datum will actually consist of two reference frames, one fixed and one timestamped. The fixed system’s coordinates will not change with time while timestamped coordinates will be able to change with time. The impact of these changes will primarily affect existing spatial datasets and precise positioning systems with absolute accuracies of less than 2 metres. For most users however, all this will be totally transparent and we can just keep using our in-car or hand-held devices as we do today.
Whether we say Happy Birthday or RIP, AGD66, its determination was a necessary and significant event in Australia’s mapping history.
Compiled by Paul Wise, February 2016
Anonymous (1947), The Sydney Morning Herald, Friday 24 January 1947, page 2 accessed at : http://trove.nla.gov.au/ndp/del/article/18018277
Ford, Reginald Arthur (1979), The Division of National Mapping’s Part in the Geodetic Survey of Australia, in The Australian Surveyor, June, September and December 1979: Volume 29, No 6, pp. 375-427; Volume 29, No 7, pp. 465-536; Volume 29, No 8, pp. 581-638, ISSN 00050326, published by the Institution of Surveyors, Australia, at : http://www.xnatmap.org/adnm/ops/prog/rafgeosvy/index.htm
Intergovernmental Committee on Surveying and Mapping (ICSM) (2015), Modernising Australia’s Datum, accessed at : http://www.icsm.gov.au/gda2020/index.html
Lambert, Bruce Phillip (1956), The National Geodetic and Topographic Survey, Annual Conference of the Institution of Engineers (Australia), Canberra, April, 1956.
Leppert, Klaus (1973), Geodesy in Australia, 1956-72, Proceedings 16th Australian Survey Congress, Canberra, pp.A1-A6, at : http://xnatmap.org/adnm/ops/prog/geod56_72.pdf