Remote Sensing Data Management for Terrestrial Ecosystem Research - an Australian Perspective
Paget, Matt1; King, Edward1; Scarth, Peter2; Johansen, Kasper3; Held, Alex1
1CSIRO Marine and Atmospheric Research, AUSTRALIA; 2Department of Science, Information Technology, Innovation and the Arts, AUSTRALIA; 3University of Queensland, AUSTRALIA

AusCover (http://data.auscover.org.au) is the remote sensing data facility of the Australian Terrestrial Ecosystem Research Network (TERN, http://www.tern.org.au). Its goal is to bring together researchers and practitioners in the satellite and airborne remote sensing sciences to develop consistent and coordinated data production, management and data delivery systems across multiple institutions. In doing so we have made remote sensing data products and associated field and metadata easy to access and understand by a wide range of terrestrial ecosystem researchers and natural resource managers and have fostered new cross-discipline activities.

Types of data produced and delivered by AusCover partners include time-series of biophysical properties (e.g. foliage projective cover, vegetation fractional cover, leaf area index, atmospheric properties), vegetation indices, fire-scars and severity, surface reflectance and BRDF corrected moderate spatial resolution imagery (e.g. Landsat, MODIS) using both global algorithms and algorithms tailored for the Australian environment, and land-cover mapping and change analyses. AusCover conducts and supports the validation of remote sensing products through the collection of field data such as systematic fractional cover surveys, ground based laser scanning for high resolution structural measurements and spectrometer measurements, while airborne hyper-spectral and LiDAR data are collected coincidently with the ground based activities to enable up-scaling to moderate spatial resolution image data sets. All these data are made freely available through the AusCover data system, subject to license conditions (CC BY by default).

The AusCover data system is somewhat unique in that it does not rely on a single point of access (a portal) but instead is a distributed network of data servers that allows replication of key data sets and both human and computer-based direct access to the data sets. A distributed but coordinated approach to data management and storage is an efficient way to manage the large archives of current and historical remote sensing data sets across a range of custodians, providers and institutions. There is a level of centralised access and discovery for metadata and visualisation and to facilitate broad links and harvesting into other disciplines and domains. Open standards have been adopted across the system from ISO compliant metadata and the promotion of non-proprietary data formats to OGC compliant web-based data delivery services.

The open and distributed data access model allows us, or anyone else, to build value-adding services on top of the data management and delivery infrastructure. We are developing a server-side subsetting and aggregating tool to allow generic space/time/parameter queries across large collections of data files with the only requirement that the files be in an open format and accessible via the OPeNDAP/THREDDS protocols. In the validation space we are partnering with the Geo-Wiki team to promote crowd-sourcing validation of remote sensing products across the Australian landscape.

The AusCover data system shares a common goal with the open data access policies of, for example, ESA and US satellite data providers and international terrestrial ecosystem and data initiatives such as the National Ecological Observatory Network, GeoLand2 and the Federation of Earth Science Information Partners. One of the early challenges for AusCover was to design a system around technologies that would allow the greatest alignment with other Australian and international open data activities at a time when development and uptake of these technologies was evolving rapidly. This will remain a challenge into the future but early design decisions have facilitated a successful and robust system that is meeting the needs of scientists, government and the community.