MF Hutchinson

Contents

• Data availability
• The First DEM of Australia
• The New DEM of Australia
• Error Analysis
• References
• Description of Data
• How to Order
• Feedback Welcome

The Australian Digital Elevation Model

The Australian Digital Elevation Model (DEM) has been developed using the package ANUDEM. It is available at a continental scale at two spatial resolutions - 1/20th degree and 1/40th degree. The DEM is also available at a finer resolution (1/60th degree or 1 minute) for individual states.

Last revision to this document:

23 July 1996

The First DEM of Australia

The first DEM of Australia at a resolution of 1/10th degree (approximately 10 km) was calculated by Moore and Simpson (1982) using the minimum curvature interpolation procedure of Briggs (1974) applied to a point elevation data set containing about 320 000 points . These data were measured during a continent-wide gravity survey conducted by the Bureau of Mineral Resources (Anfiloff et al., 1976). Though detailed enough to detect significant geological structures (Harrington et al., 1982), this DEM suffered from a number of acknowledged limitations from the point of view of hydrological analysis. Extreme heights were not well represented and the 1/10th degree resolution could only support quite generalized drainage structures. The data were also known to contain a number of errors.

The New DEM of Australia

The new 1/40th degree DEM of Australia has been calculated by the elevation specific gridding technique described by Hutchinson (1988, 1989). This technique ensures that the DEM has a connected drainage structure by automatically removing spurious pits or sinks. The DEM also incorporates the stream line network digitized from the 1:2.5 M scale map of Australia (Division of National Mapping, 1984). A summary of the point elevation and stream line data used is given in Table I. A total of about 580 000 elevation data points were used, of which 400 000 were provided by the Bureau of Mineral Resources. These included the 320 000 elevation points used by Moore and Simpson (1982). The inclusion of the trigonometric data points ensured that most of the principal peaks were incorporated into the DEM. Additional point and stream line data were digitized from 1:250 000 scale topographic maps to provide more detail and to remove remaining drainage anomalies, particularly in areas with complex terrain. Sink data points were also digitized to prevent drainage clearance of genuine depressions, although the drainage enforcement algorithm is generally sufficiently sensitive to ensure that genuine depressions are not cleared, whether or not they have been identified as such in the data (Hutchinson, 1989).

Error Analysis

Numerous large errors in all of the data sets in Table I were detected and corrected by examining remaining sinks in the fitted DEM. Elevation errors as small as 20m were easily detected in this way. This was consistent with the estimated errors quoted in Table I and the setting of the first elevation tolerance in the gridding procedure to 15m (Hutchinson, 1989). Remaining sinks also indicated areas where more point or stream line data were required to properly define drainage structure. Further details of the methods used and the technical problems which had to be overcome are described in Hutchinson and Dowling (1991). In particular, positional errors in the stream line data digitized from the 1:2.5 M scale map were reduced to the magnitude quoted in Table I by ' rubber sheeting' the stream line data using a set of control points comprising one point for each 1/2 degree cell across Australia. The control points were digitized from 1:250 000 scale topographic maps. It should be noted that the DEM data are subject to ongoing revision and augmentation. The consistency of the drainage analysis shown in Figure 1 with existing analyses indicates that the remaining errors in the DEM data are relatively minor.

The overall accuracy of the DEM is dependent on the local relief of the actual landscape and the resolution of the DEM, as well as the accuracy of the elevation and stream line data. In areas with low relief elevation errors in the DEM approach elevation errors in the data of about 10 m. In areas with complex terrain elevation errors may be well in excess of 100m. These occur when the gridding procedure has to resolve conflicts between maintaining fidelity to point elevation data and maintaining descent down stream lines at the chosen grid resolution. All such conflicts were resolved in favour of the stream line data. Derived terrain parameters such as slope must be regarded as being highly generalized in these areas.

Table I. Sources and estimated maximum errors of the point elevation, stream line and sink point data used to construct the 1/40th degree DEM of Australia. The numbers of points in the stream line and coastline data were determined after the data had been generalized to the resolution of the DEM. BMR denotes the Bureau of Mineral Resources. Data marked by * were provided by the Division of National Mapping.

____________________________________________________________________________
Data source                Number     Horizontal    Vertical
                          of points     error         error
                                         (km)          (m)
____________________________________________________________________________


Point Elevations
  BMR ground survey        400 000        0.1           7
  Trigonometric points*     19 000        0.1           1
  Bench marks*              83 000        0.5          10
Coastline (1:250 000)*      10 000        0.1           0
  1:250 000 scale maps      65 000        0.1          10
Stream Lines
  1:2 500 000 scale map*    15 000        1.0           -
  1:250 000 scale maps       7 000        0.1           -
Sink Points
  1:250 000 scale maps         400        0.1          10

_________________________________________________________________________
  

References

Anfiloff,W.,Barlow,B.C.,Murray,A.S.,Denham,D.,and Sandford,R. 1976. 'Compilation and production of the 1976 Gravity Map of Australia'.Bureau of Mineral Resources J. Austral. Geol. Geophys. 1: 273-276.

Briggs,I.C. 1974. 'Machine contouring using minimum curvature'. Geophysics 39: 39-48.

Division of National Mapping (NATMAP) 1984. MAP R201 - Geographic Map of Australia, Division of National Mapping, Department of Resources and Energy, Canberra.

Harrington, H.J.,Simpson,C.J.,and Moore,R.F. 1982. 'Analysis of continental structures using a digital terrain model (DTM) of Australia'. Bureau of Mineral Resources J. Austral. Geol. Geophys. 7: 68-72.

Hutchinson,M.F. 1988. 'Calculation of hydrologically sound digital elevation models', Third International Symposium on Spatial Data Handling, Sydney ,International Geographical Union, Columbus, 117-133.

Hutchinson,M.F. 1989. 'A new procedure for gridding elevation and stream line data with automatic removal of spurious pits', J. Hydrol. 106: 211-232.

Hutchinson,M.F. and Dowling,T.I. 1991. A continental hydrological assessment of a new grid-based digital elevation model of Australia'. Hydrological Processes 5: 45-58.

Moore,R.F.and Simpson,C.J. 1982.'Computer manipulation of a digital terrain model (DTM) of Australia'.Bureau Mineral Resources J. Austral. Geol. Geophys. 7: 63-67.

Description of Data

The Australian digital elevation model is available at two spatial resolutions. The details of the two DEMs are as follows:-

The grids are written by rows starting at the south-western corner of the grid long. 112(E) and -44 (S) and moving east along each row.

Also available in ARC/INFO ASCIIGRID form.

Feedback Welcome

Your comments and feedback on the documentation and the data itself are invaluable to us.

Please click here to mail your comments to Michael Hutchinson.