News

TERN’s national ecosystem data infrastructure: an overview

The ecosystem science and management research cycle

Associate Science Directors Stuart Phinn and Andy Lowe explain how TERN’s national ecosystem data infrastructure is delivering efficiency gains, increasing effectiveness and increasing integration across disciplines for Australia’s ecosystem scientists and managers. TERN’s infrastructure is catalysing better returns on investment in ecosystem science for Australia.

TERN is working to bring about an Australian ecosystem science community that has undergone transformational change – from one in which effort was frequently fragmented, inefficient and short-term, to one that is national, networked and delivering for Australia’s future. Key to this vision is a coordinated national capacity for data collection, management, licensing, storage, and subsequent discovery, which TERN is providing through its national ecosystem data infrastructure. Our philosophy is ‘collect data once – make it discoverable – use it many times’.

A typical research cycle informing ecosystem science and management (including the processes that researchers often work through) is shown in the figure below. TERN is engaged at all points of this cycle. By building on significant past investments in ecosystem data collection, and working closely with a diverse range of partners, we’ve generated specific tools and infrastructure that can help increase the efficiency and effectiveness with which Australia’s ecosystem science community deals with data and advances knowledge.

Let’s look at the process step by step. The identification of important gaps in our ecosystem science knowledge, and/or our monitoring and management approaches, leads to the formulation of specific research questions. Already, at the proposal and planning stage, efficiency gains are obtained through improved accessibility to long-term ecological and biophysical datasets, the infrastructure used to collect these, and standardised national methods that can be selected from the ecosystem science toolbox. Examples include the national standards for ecological monitoring currently being developed through TERN’s Multi-Scale Plot Network, and the Australian Supersite Network’s approaches to handling and delivering ecosystem data streams. There’s no need to spend time re-inventing a method if there’s one available on the TERN shelf that will do the job, or if a quick search reveals that there are guidelines for best practice or existing long-term datasets that can be built on.

Many TERN facilities are engaged in making the next steps – data collection, verification, quality assurance and control – more efficient for researchers. Examples include improved access to national data collection infrastructure that’s been built and maintained over the past three decades; our diverse range of mobile field-data-collection apps, such as the one being supported through the TERN Soils facility, that help to minimise the potential for error in data collection and entry, and provide seamless integration into storage infrastructure; AusCover’s validations of remote-sensing algorithms for Australian conditions; and the four levels of quality assurance and validation that OzFlux has developed to manage real-time gas flux data streams effectively.

TERN encourages the sharing of data among Australian ecosystem scientists, but this is only possible when metadata descriptions are accurate and complete, data contributors receive appropriate recognition for their work, and when data users understand, in advance, the responsibilities placed upon them when they use datasets collected by others. TERN has worked hard with the broad range of scientific communities that make up ecosystem sciences in Australia to put the necessary integrative data infrastructure in place. For example, all TERN facilities promote the use of internationally accepted metadata standards across different ecosystem science disciplines, and the central portal uses Australia’s most common data interchange format.

Our new DOI (digital object identifier)-minting service is available to provide persistent identifier to datasets published by TERN, making them citable. This supports the increasing perception of datasets as scholarly contributions in themselves, and improves accessibility and reusability. We have also developed the TERN data licensing framework, including a suite of recommended and flexible data licenses appropriate for ecosystem datasets under a range of circumstances.

The next step in the process calls for safe long-term storage and preservation of the hard-won datasets, and their associated metadata, such that they are easily discoverable by potential future users. TERN’s Australian Coastal Ecosystems Facility is doing great work in bringing together previously dispersed, hard-to-find datasets onto a common platform. Elsewhere in this newsletter you can read about TERN’s new Data Discovery Portal, which has gone live this month, and which for the first time permits metadata searching across multiple diverse ecosystem science disciplines, and links through to actual datasets wherever feasible. The potential efficiency gains are obvious.

The Australian Ecological Knowledge and Observation System (ÆKOS) is one of the most exciting advances in ecosystem data infrastructure achieved under the TERN banner. Developed through TERN’s Eco-informatics facility, AEKOS’s world-leading semantic approach not only permits storage, preservation and access to the full richness of ecological plot data, including collection methodologies and photos, it also promotes the integration of heterogeneous datasets in ways that were not previously possible. The AEKOS system now houses a large number of critical state and territory government and researcher survey and monitoring datasets, and the long-term plot data that forms part of the MSPN is currently being uploaded. We look forward to the trial beta release of AEKOS in December 2012.

The next step in the process requires data analysis, integration and synthesis, and here too TERN’s approach to providing critical research infrastructure has been innovative. Two TERN facilities deserve special mention in this context. Ecological Modelling and Scaling Infrastructure (e-MAST) brings together some of the datasets now flowing freely from other TERN facilities to test and improve models used for such applications as future climate scenarios and the assessment of primary production. The Australian Centre for Ecological Analysis and Synthesis (ACEAS) provides direct links between scientists and policymakers in ways that were not previously possible in this country. Sharing of data and collaborative development of ideas and solutions in this context helps address fundamental issues in ecology and allied fields, and generates research outputs that are explicitly relevant for policy.

The effectiveness and efficiency dividends that accrue from the use of TERN’s data infrastructure through all of these steps then provide the potential for new kinds of complex research outputs that cross disciplinary and institutional divides. The recent release of Persistent Green, a new Australian remote sensing product that has been generated through cooperation and collaboration between many stakeholders with the support of TERN’s AusCover Facility, is an excellent example.

In summary, TERN’s national ecosystem data infrastructure delivers:

  • efficiency gains for researchers – using TERN’s infrastructure they can now do more with less (methodological standardisation, tools and products to improve data collection and quality assurance processes, and centralised portals to store data and search for relevant existing datasets, which can then be downloaded often under open licensing arrangements);
  • productivity gains for researchers – the existence of TERN’s infrastructure provides large volume of pre-existing datasets for researchers to address complex questions, in new ways, resulting in new knowledge (a data discovery portal providing easy access to quality data by permitting searching across disciplines, new capacities for analysing and synthesising ecosystem data, and increased likelihood of policy-relevant outcomes);
  • increased recognition of research effort – TERN’s infrastructure is the foundation for open access to ecosystem data in Australia, because it provides incentives to share and re-use of data: the capacity for semantic data storage, provision of DOI-minting services and availability of appropriate licensing, in a global climate of increasing recognition of data publication.

In this way, more and higher-quality results can be achieved from single investments in research and data collection. TERN’s national ecosystem data infrastructure is therefore enabling the Australian community as a whole to realise a better return on public investment in ecosystem science.

Our successful engagement with many partners – from data collection to analysis, modelling and management, across Australia’s multiple ecosystem science communities, through state and federal government, to universities, industries, and community groups – means our approach represents a significant advance upon previous large-scale ecosystem data life-cycle initiatives. A range of stakeholders are now beginning to benefit from the efficiency and productivity gains now feasible thanks to the critical data infrastructure being provided through TERN. Some of their stories will be told at the special session on TERN’s national ecosystem data infrastructure at the 2013 TERN Symposium in Canberra in February (registration is free and open now!).

Published in the TERN e-Newsletter, October 2012

Share Article