Published works

Up to date information about the spatial distribution of water that is suitable for use by firefighting aircraft is required to:

• Provide pilots with pre-flight information about the availability of water close to the fire ground.  This allows pilots to source water more quickly and therefore reach the fire ground with water more rapidly, and minimize the travel time between pick-up and drop when multiple drops are required.
• Contribute to information used by fire authorities to optimize the distribution of their firefighting aircraft throughout the firefighting season, to optimize the efficiency of the aerial firefighting fleet.

The Digital Earth Australia (DEA) Waterbodies tool provides information water in waterbodies bigger than four Olympic swimming pools.  This information is updated with every cloud free satellite overpass (satellite revisit every 16 days).

The aim of this project was to:

• Evaluate the extent to which the current DEA Waterbodies product can be used to provide both start-of-season and near-real-time pre-flight information.
• Deliver a proof-of-concept web feature service that includes up-to-date fill levels and waterbody suitability for different types of firefighting aircraft and to ensure that the web service could be accessed via NAFC’s Arena spatial platform.
• Identify improvements to the DEA Waterbodies product that would be required to further increase its utility to the National Aerial Firefighting Centre, state and territory firefighting agencies and aircraft operators.

FINDINGS

DEA Waterbodies v1 was found to be suitable for start-of-season, severe fire danger period and start-of-fire information for command centres seeking to identify water availability within range of Helitak airpads (both permanent and temporary airpads).  However, at present, its utility for pre-flight information for pilots was minimal based on the following limitations.

Waterbody size limitations

Limitations of DEA Waterbodies v1 Noting that the 25 metre pixel size provides a fundamental constraint on the minimum size of waterbodies that can be reliably detected using Landsat. A waterbody must be larger than 3125m2 (5 Landsat pixels) to be included in the DEA Waterbodies v1 polygon set.  The water surface area in these polygons will be tracked down to 625m2 (a single Landsat pixel). This means that there will be many waterbodies that are suitable for access by small rotary wing aircraft that are not captured in the DEA Waterbodies v1 product.  A future version of the DEA Waterbodies product could include smaller waterbodies to improve its utility for aerial firefighting stakeholders.  This could be achieved by:

• relaxing the minimum size constraint of DEA waterbodies (capture waterbodies as small as a single pixel), and/or
• applying the DEA Waterbodies workflow to Sentinel 2 imagery (10 metre pixel size).

Timeliness limitations

Waterbody fill level information is of greatest utility to firefighting agencies when it is as up-to-date as possible.  The timeliness of DEA waterbodies fill level information is determined by two elements:

• Latency (the time between when a waterbody is observed by satellite and when the fill level information is available in the firefighting agencies spatial information system i.e. Arena for NAFC or EMAP for Victorian CFA).  This is determined by the level of automation of the IT systems that sit between the satellite itself, and the DEA waterbodies web service.
• Observation frequency (how frequently a satellite observes that waterbody).  This is determined by the number of satellites operating and the presence/absence of cloud/smoke/haze.  At the time of writing the observation frequency for DEA Waterbodies is ~16 days (Landsat 8 only), but has the potential to be ~ 3 days (Landsat 8 and 9, Sentinel 2a and 2b).

Polygon attributes

Waterbody fill level information is of greatest utility to firefighting agencies when presented in the context of whether that waterbody is suitable for use by a specific aircraft type.  Waterbody-aircraft suitability can be determined either via matching the historical fill activity for a particular aircraft type (demonstrated in this study) or via a spatial analysis of constraints (workflow described here for future implementation).