Collecting Nationally Important Spectral Datasets

Geoscience Australia (GA) is developing a validation strategy for the products generated by the Digital Earth Australia (DEA) Program. As the baseline dataset for a number of derivative products, validation of the DEA Surface Reflectance (NBAR) dataset is being given priority. Field-based validation will be used to establish uncertainty in the DEA Reflectance product and will be critical in defining fitness for purpose.

To achieve this outcome, two significant and complementary activities have commenced:

  1. CSIRO and some universities are assisting GA in the collection of new field reflectance data at selected sites contemporaneously with satellite overpasses, in a nationally coordinated approach to collect field validation data.
  2. Building on the work done on the SPECCHIO spectral database at the University of Wollongong, GA will host a national spectral data repository, and will conduct a national audit of existing laboratory and field-derived spectral data with a view to incorporating suitable data into the repository.

In support of both these activities, we are looking to include any existing spectral libraries that may help compile quality data including:

  1. End-member data that is supported by appropriate metadata (site name, site description, target description, photos…). This could include reflectance and associated (e.g. transmission, absorption if appropriate) spectra of suitable targets (soils, vegetation, corals, minerals), habitat types and species.
  2. Historical field validation data – this would consist of field / airborne data that could potentially be applied to validation of continental scale surface reflectance products based on Landsat 5, 7, 8 and Sentinel-2 data. Ideally, the field data for validation of surface reflectance products would have been collected as part of a campaign synchronised with satellite and / or airborne data acquisition and have appropriate metadata linked to it and be able to be openly shared with appropriate acknowledgement.

We are therefore seeking the Australian EO community’s help in identifying currently available spectral datasets fitting the above two categories to help create a valuable and enduring resource for the Australian Earth observation community.

If you hold any field data, or aware of data that you believe could potentially be used for validation of EO products, please contact Tim Malthus ( and Medhavy Thankappan  (

Tim Malthus (CSIRO) and Medhavy Thankappan (GA)