Why are there differences between the new GDA2020 Australian coordinate system and the previous MGA9


 Product:MicroStation
 Version:V8i SS10, CONNECT
 Environment:N\A
 Area:Coordinate Systems
 Subarea:Geographic

Problem

We have project data in the Australian Project GCS of MGA94 Zone 56 and are finding we are getting 3 different results with Bing Maps Aerial background display and data reprojected from other coordinate systems. Between the legacy MGA94-56, MGA94-56-GSB and 7P we get different results every time. Isn't the data supposed to have identical results seeing that they are the same coordinate system, and we are not technically reprojecting anything?

Solution

There are three versions of MGA94 zoning systems because these are based on three different datums which implement 3 different datum transformations to other datums.

The older MGA94-56 is based on datum GDA94 which is considered coincident with WGS84, but it was coincident with the older variant of WGS84 many years ago. Unfortunately the Australian sub-continent moves at a relatively high speed as goes for tectonic plates. This implies that the reference system must be modernized periodically. This led to the creation of the GDA2020 datum which is the current datum and this datum is considered coincident with the current version of WGS84. Normally all control points in Australia should be published based on this datum and newer project should probably be based on a GCS based on this datum (MGA2020-56 for example).

All GDA94 based GCS are useful for older projects but care must be taken to understand when using it how to use it and what datum transformation ought to be applied.

MGA94-56 is based on datum GDA94 which considers GDA94 datum coincident with WGS84 (but remember it is for the older version of WGS84)

MGA94-7P-56 is based on GDA94-7P which implements a Positional Vector transformation (7-Parameter) to GDA2020 (which is now coincident to the newer WGS84). This version is an approximation provided by the Australian government for software that do not support the normative method based on NTv2 files. This version is only provided to correct data that was previously geolocated using 7 parameter. It should not be used otherwise.

MGA94/GSB-56 is the version that implement the datum transformation to GDA2020 using the NTv2 grid file method. This is the recommended method to convert from GDA94 to GDA2020. Since GDA2020 is now considered coincident to WGS84 the datum transformation through NTv2 files will also apply when converting to WGS84.

So these are the reasons why we have three different conversion to WGS84.

Now the question is what does Bing Map uses as coordinate foundation since they use the Web Mercator projection supposedly based on WGS84. Usually Bing tends to use the most modern local datum meaning it ought to be the GDA2020 version but it could still be on the GDA94 version of WGS84. Obtaining a better match using either assumption is not a statement of the exact datum used by Bing. I never noted a statement of accuracy for Bing imagery but by experience with previous client data, Bing is positioned within an accuracy ranging from 30 centimeters to over 70 meters depending on the part of the world. Urban areas in the United States of America apparently have the best accuracy while distant rural or uninhabited areas have poor accuracy. I believe it would be wrong to assume Bing Imagery to be accurately positioned to better than 50 centimeters in Australia.

So to summarize to a short answer:

See also

Where can I see the new Australian Geographic Coordinate System GDA2020

GDA Frequently Asked Questions | Intergovernmental Committee on Surveying and Mapping (icsm.gov.au)

Other language sources

 Original Author:Bentley Technical Support Group