Landsat Image Based Temporal and Spatial Analysis of Farm Dams in Western Victoria

With the continuing pressure on Australia’s water resources and interand intra-annual rainfall variability, farmers, especially in rainfed agricultural and pastoral zones, have been developing farm dams as a mean of providing additional water sources for irrigation and stock. The density of farm dams has increased over time, and an accurate knowledge of the spatial distribution of farm dams is essential for water resource management at the catchment level. There are more than 2 million farm dams across Australia, most with a storage area of <5-10 ha. These small dams are frequently unlicensed. In Victoria stock and domestic dams do not require a license, so more than 90% of farm dams, contributing 85% to the total storage capacity, are not documented, despite their significant impact on water resources and stream flow. To document farm dam distribution and historical trends in their development, an approach has been developed using medium resolution Landsat images coupled with Google Earth images. This approach is cost-effective, time-efficient, reliable, accurate and repeatable. Various image manipulation/enhancement and analysis techniques were evaluated, and showed that even small water bodies can be identified using False Colour Composites further enhanced by using image enhancement and transformation techniques and combined with the band 5/4 ratio. Density slicing of the mid-infrared band (5) was not useful for very small farm dams as the water reflectance is highly variable due to available water quantity and quality. Using the Landsat data for 1973 to 2004 for a small area in western Victoria, 99% of farm dams identified using the high resolution Google Earth images were successfully located, showing that Landsat data is a good compromise for this purpose, especially considering the cost. Over the 21 years from 1973 to 1993, there was a 276% increase in farm dams, driven by a dramatic increase in livestock population, particularly cattle, as the dams are mainly used for stock watering. The serious droughts of 1982-1983 and 1997-2010 had little influence on the rate of increase of farm dam numbers. After 1993 there was a much slower rate of dam construction, due to the fact that most suitable dam sites had already been utilized, and there was a shift from dryland pastures to dryland cropping, with a concomitant decrease in demand for watering points.