Science and Strategy

The Science and Strategy Section provides strategic science leadership and advice related to Earth observation from space. The multi-disciplinary team of remote sensing professionals in the section apply scientific methods for the calibration of satellite observations and develop techniques for the analysis and application of Earth observation data to support government priorities.

The Section's major projects are:

Land cover mapping

Land cover is the observed biophysical cover on the Earth’s surface, including native vegetation, soils, exposed rocks and water bodies as well as anthropogenic elements such as plantations, crops and built environments. Different land cover types display distinct responses due to seasonal, climatic and anthropogenic drivers. Classifying these responses that have been recorded by satellites over a number of years provides a robust and repeatable way of characterising land cover types.

The land cover mapping project will deliver the first nationally consistent and thematically comprehensive land cover reference for Australia and is the result of collaboration between Geoscience Australia and the Australian Bureau of Agriculture and Resource Economics – Bureau of Rural Sciences. Outputs from this project will provide a national baseline for land cover analyses, and enable Australia to meet many environmental reporting needs. The land cover mapping capability will play a crucial role in informing issues of national significance such as sustainable farming practices, management of water resources, air quality, soil erosion, forest management, as well as emergency management.

Detection of hotspots using time-series images from geostationary satellites

Early detection of bushfires using remote sensing techniques provides critical information for emergency management agencies. Detecting fire using remote sensing is based on thermal anomaly or ‘hotspot’. This technique identifies areas that are ‘hotter’ than a certain temperature threshold and flags them as hotspots. The Sentinel bushfire monitoring system uses hotspots information derived from orbital rather than geostationary satellites, meaning that it would be more than four hours between subsequent hotspots observations. Outputs from the geostationary hotspot detection project will greatly reduce the lead-time for bush fire detection from five to six hours to one hour. Time series analysis tools developed for this project will increase the accuracy of hotspot detection and reduce the rate of false alarms.

Detection and characterisation of offshore natural oil seeps

This project develops and improves the methodology for mapping offshore oil seeps using radar and optical Earth observation data. Outputs from this project have been adopted for cost-effective screening of offshore petroleum basins to support the Australian Government’s Offshore Energy Security Program. This project has developed a semi-automated method for screening hundreds of radar images for the presence of natural oil seeps throughout the Australian marine jurisdiction, obviating the need for analyst intervention. This project also characterises the optical properties of Australian oils in the laboratory to determine the feasibility of using satellite based optical data to detect these oils in the marine environment. The outputs provide guidance for investment decisions associated with the use of optical satellite imagery.

Shallow-water depth estimation

The extraction of depth information from Earth observation data is an emerging capability and has potential applications in environmental monitoring and risk assessment studies nationally and in the region. Cost-effective application of the methodology for mapping shallow water bathymetry and benthic environments over extensive areas such as the Great Barrier Reef has been demonstrated with data from the Advanced Land Observing Satellite (ALOS). This project developed a satellite-based high resolution bathymetry map of Lord Howe Rise; unique spectral information collected through a field campaign at Lord Howe will enhance the depth maps for Lord Howe. Further research is being planned to strengthen this emerging capability.

Standardising corrections to Earth observation data

Correcting for atmospheric and land surface bidirectional reflectance distribution function (BRDF) effects is particularly important for standardising time series Earth observation data. Standard corrections remove atmospheric and BRDF influences to provide a more accurate measure of surface reflectance, enabling the comparison of data from multiple sensors and time periods. This project focuses on the development of operational procedures to correct higher resolution sensor data (Landsat and similar resolution). The approach uses combined physics-based models of the atmosphere and BRDF. The method has been tested using Landsat data for two sites with different land covers in Australia; the retrieved surface reflectance values were in good agreement with ground based spectroradiometer measurements. The outputs from this project will benefit a number of applications that use time series Earth observation data such as land cover mapping.

Topic contact: Last updated: November 18, 2010