Using drones to measure forest water use

Around 80 percent of Melbourne’s drinking water is sourced from uninhabited forests in the Yarra Ranges and Central Highlands. More than 1,570 square kilometres is reserved for water catchment. This streamflow has been a reliable source of water for well over 130 years. As a water supplier dependent on streamflow, it is important that Melbourne Water understands how streamflow varies over time.


For a basic estimation of streamflow from forested catchments, we measure rainfall minus the water used by the forest and the water lost due to evaporation. This water use and water loss is collectively referred to as evapotranspiration, or ET.

Accurately determining ET is difficult. Traditional methods are referred to as ‘bottom up’ and use historical data to estimate ET, which is then used to predict future ET.

The problem with this approach, is that future ET is assumed to be stable over time, which is not true. For example, the Mountain Ash forests that supply much of Melbourne’s water are well known for changing their water use as they age. Other issues such as fires and climate change will also alter ET over time.

A ‘top down’ approach estimates streamflow by measuring forest water use and subtracting it from rainfall. This means variability in ET can be directly incorporated in the streamflow assessment. But how do you accurately measure forest water use?

One method is known as the heat pulse method where a small probe is inserted into each tree to determine the amount of water used. While this is suitable in small forest plots, it is not practical on a larger scale.

Drones, on the other hand, are useful for measuring larger-scale forests. They capture a forest’s characteristics allowing water use to be estimated using this information. As a stand-alone method, however, it is not entirely reliable. But researchers at the University of Newcastle have developed an additional element to this approach: by attaching a small, remote sensing LiDAR instrument to the drone, precise, three-dimensional information on various tree parameters is generated. This can then be used to determine ET directly.

Innovations such as these enable Melbourne Water to more accurately forecast future streamflow, in turn helping us make more timely decisions about Melbourne’s drinking water supply.

Last updated:
20 May 2019