ANU Kioloa/NASA Pathfinder GLCTS Site (1997): P.I./Data Custodian; Dr. Brian Lees

Kioloa DTM (1997 version).

The DTM was derived from digitizing 1:25000 topographic maps.
The maps were cut into tiles so that each tile covered an area
of 5.25km x 5.25km (175 x 175 cells). Students digitized and
interpolated the tiles which were then edited by  Belinda
Allison. The tiles were pasted together to produce a single
image. 

filename: dtmfilt.asc
datatype: real
filetype: ascii
columns: 1925
rows: 2450
units: metres
cellsize: 30m
min x: 229000
max x: 286750
min y: 6043000
max y: 6116500
min value: 0
max value: 1120

Lineage:

The DTM was digitized from the 1:25000 Topographic map series
(1984 edition) produced by the Central Mapping Authority of
NSW. They have a contour interval of 10m and are based on a
UTM projection. The individual tiles were digitized, rasterised
and interpolated in 1993, 1994 and 1995 and then edited and
joined together in 1996. A mask was created by digitizing the 
coastline. The coastline was rasterised giving a grid of exactly 
the same size as the DTM with each cell having a value of 1 or 0; 
1 denoting 'land' and 0 for 'water'. The dtm and the ocean mask were 
multiplied together resulting in a DTM with zero values for the ocean.
A 3x3 mean filter was applied to the image. 

The interpolation algorithm used was IDRISI INTERCON v3.02. 

Completeness:

The DTM is complete. Zero values have been given to the ocean
to give a complete rectangle of values 73.5km x 57.75km in size.

Consistency:

Consistancy is variable as a large number of students were
involved in the digitizing. However each tile was checked and 
corrected by Brian Lees and Belinda Allison.

Correctness:

Correctness is variable. As much as possible erroneous contour
values were corrected and leaks between rasterised contours
were removed by updating the value of the cells. The cliffs
in the north-east of the dataset were poorly defined by contours
and remain approximate. The error of this DTM
is discussed by Van Neil et al., (2004). 

"The mean of the DEM error assessed against the GPS data was 5.8 m, showing that the DEM overestimates elevation.  The errors are comparable to DEM error levels found in USGS and British Ordnance Survey DEMs.  The error was also shown to be normally distributed and to be slightly spatially correlated at 30 m intervals. "