546
SCANSAR data. Eight images annually will be
calibrated, registered together and interrogated in
an attempt to derive a map of the distribution of
wetlands across Northern Australia. Comparison
of wet and dry season images will also be
undertaken in an attempt to try and distinguish
between tidal and freshwater ^ wetland
environments.
Imagery from at least four contiguous swaths
(depending on orbital paths) will be needed to map
the full extent of wetlands across Northern
Australia.
Regional mapping of the wetlands of Northern
Australia will be carried out in all three years of
the project.
Detailed mapping of the Alligator Rivers Region
wetland environments using RADARSAT data.
Hydrologic, geomorphic and ecological field and
reference data is being acquired after the end of
the wet season to coincide with the recession of
floodwaters. Successive stages in the drying of the
floodplain will be monitored and the changing
environmental conditions identified and compared
to changes in the backscatter return recorded on
successive images.
Closely related to (2) will be the determination of
specific wetland processes associated with:
(a) hydrology - monitoring the seasonal
pattern of floodwater recession;
measurement of sediment and biological
loads.
(b) geomorphic changes - examining the
interface between freshwater and tidal
estuary areas over a 6-9 month period to
determine if headwater extension of
saltwater streams is occurring into
freshwater floodplains.
(c) characterisation of vegetation found
within the
(1) tidal mangroves
(ii) coastal salt marshes and mud flats
(iii) seasonally inundated floodplain,
lakes and waterholes
(iv) | macrophytic distributions
(v) adjacent abutting
woodlands and forests
upland
Regional mapping and the monitoring of
hydrologic, geomorphic and ecological patterns
and processes will be repeated and extended in
1999. The timing and sequence of RADARSAT
data requests will be adjusted based on the spatial
and temporal patterns observed in the landscape in
1998.
Also, in 1998, the research will focus on the
characterisation, field sampling and the
measurement of biomass in selected vegetation
habitats as well as determining changes in the
pattern of inundation and recession of floodwaters.
S: A GIS relational database is being established at
the Environmental Research Institute of the
Supervising Scientist (ERISS) in Kakadu National
Park to incorporate the results of the ongoing
hydrological, geomorphic, ecological ^ and
biochemical studies related to the management of
wetlands. Image processing and analysis of
RADARSAT will be undertaken at the Centre for
Remote Sensing and GIS at the University of New
south Wales and transferred to the ERISS
database.
6. A regional map of the wetlands of Northern
Australia will be published incorporating
RADARSAT data and planimetric information
derived from other sources.
Sequential maps of floodplain inundation levels
for the Alligator Rivers Region for 1998 and 1999
will be assessed and absolute and relative changes
in the hydrologic, geomorphic and ecological
characteristics of the wetland environment
examined and described. Strategies for assessing
and verifying longer term changes in the wetland
environment will be developed.
The methodology developed for producing the
sequential mapping of individual wetland
catchments in the Alligator Rivers Region will be
applied to the Ord River drainage basin in Western
Australia and other ecological catchments across
Northern Australia.
RESULTS TO DATE
The results of the analysis of the first year of
RADARSAT data for examining and describing the
vegetation response associated with the wet-dry seasonal
conditions will be presented at this conference, however,
significant results have been made in several related
studies. The first empirically based statistical
relationships between radar backscatter and = field
measurements of selected tree parameters in tropical
Australian forests that demonstrate multi-polarimetric
SAR responds to vegetation structure (eg. total biomass,
leaf-area index, branch surface/volume ratio etc.) in
predictable ways have been demonstrated (Imhoff, et al,
1997).
These results show radar sensitivity to floristically
induced structural changes across the Kakadu study site ,
indicating that given appropriate algorithms,
physiognomic classifications of trees and community
typology may be possible with radar.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998