Research staff

Grant Williamson
Post-Doctoral Research Associate
Phone: (08) 8946 7726
Email: grant.williamson@cdu.edu.au

Drainage works constructed for mosquito control purposes have had a remarkable impact on the vegetation and ecology of Leanyer swamp, converting a fresh and brackish water swamp complex first to bare saltpans, then to rapidly encroaching mangroves. The mangroves have moved up to 1km inland from their prior distribution along the drains.

Over the period of the study of landscape change, the coastal swamps of Darwin have experienced a number of other influences besides man-made drainage works. Most notably, Cyclone Tracy flattened extensive areas of the mangroves. There are reports of sea-level rise, which may influence vegetation presence and patterns in this extremely flat intertidal zone; the distance between the lowest tide and the highest tide is over 5km. And urbanisation and climate change have had the potential to increase the flow of freshwater into the upper reaches of the swamps.

Leanyer drains

Grant is analysing vegetation change using aerial photographs and satellite images spanning the last 30 years, attempting to correlate vegetation communities through time with these drivers, in order to determine how much change in the vegetation in the swamps is related to direct human activity, and how much may act as an indicator of broader climate factors.

One important data layer required is elevation. Very slight differences in elevation can have a dramatic influence on vegetation type, tidal access and hydrology. Unfortunately, the 90sec DEM has neither the vertical nor horizontal accuracy to be useful in this system, while the DPI's contour maps also lack the vertical resolution to be of use.  Grant has taken advantage of the new Magellan Promark 3 differential GPS system to conduct an elevation survey in the swamp, in order to fill in the gaps in the DPI's map. Despite a steep learning curve, this appears to have been successful, with elevations recorded at sub-metre accuracy post-processed against the Darwin base-station, and with close to centimetre accuracy using the DGPS base-station.

Vegetation classification for the study has been done manually, by visually categorising randomly placed points. Automatic classification was not used due to the extremely variable nature of the aerial photography available, which included monochrome, colour-infrared, and colour images, varying in scale from 1:5000 to 1:50000, taken at different times of the year. All that remains to do is finish a few thousand more point classifications and to construct the statistical model.