Improving knowledge transfer to support Australian Marine Park decision making and management effectiveness evaluation
Parks Australia requires data products that are “fit for purpose” in order to effectively manage Australia’s marine parks. The first objective of this project is to assist Parks Australia develop and implement a data product quality assurance framework, supported by an Australia Marine Data Commons, by (i) identifying the use cases and data products it needs to answer key management questions; (ii) developing a suite of quality assessment criteria; (iii) assessing existing data products and their associated repositories against these criteria; and (iv) document recommendations that will help ensure data products become fit for their intended purpose.
A data product is a re-useable data asset designed to support one or more data use cases. A data use case is a description of a practical application of a data product. A use case typically describes how a user performs a function, for example how a user answers a key management question with a data product. Prior to this project Parks Australia identified six key questions (KQ’s) that define their most important data use cases.
1. What natural values are in the area of interest?
2. What state or condition are the natural values in?
3. How special are the natural values?
4. How might the natural values respond to a pressure, activity, or management action?
5. Where is the activity, incident, or pressure occurring in relation to where the values are?
6. What management actions can be taken to contribute to zone objectives and conservation goals for the natural values?
There are other management questions that park managers need to answer but these six questions were considered to be a priority and at an appropriate level of detail for the purposes of this project.
Prior to the project Parks Australia held two workshops to identify a list of data products, organised into a data product hierarchy, that the project might assess. By examining the workflows and decisions Parks Australia staff use when answering key questions 1 and 3, as well as 2 and 5, the project was able to identify 20 priority data products and associated repositories from among the 90 plus data products identified in these workshops.
The project was unable to describe the use cases for key questions 4 and 6 because Parks Australia has yet to establish workflows that identify the data products that they will use to answer these two questions. These questions therefore will require additional analysis that accounts for Parks Australia’s management approach.
In the workshops conducted prior to the project Parks staff also identified a range of quality concerns associated with data products. These concerns were categorised and organised into an initial data product quality assessment framework addressing 6 aspects of data quality: Requirements, Science, Production, Stewardship, Service and Use.
This framework provided the basis for a more detailed data product quality assessment undertaken in this project. For each quality aspect in the framework the project developed series of questions which were used to test if the product was “fit for purpose’ and elicit more detailed quality concerns.
The quality evaluation was performed using a questionnaire style survey addressing all six quality criteria. There was a total of 41 questions, 4 addressing the Requirements Quality criteria, 7 directed to the Science Quality criteria, 10 under Production Quality, 9 under Stewardship Quality, 7 addressing the Service Quality criteria, and 4 directed to the Use Quality criteria. These questions were answered by staff at Parks Australia and the project team. Answers provided for Production quality, Stewardship quality and Service quality were subsequently checked by conducting interviews with the data product’s owner, manager or representative.
Each of the data quality assessment questions were designed for three primary responses: “Yes”, “No”, or “Partially”. The overall quality of a data product was then based on how many of the quality assurance criteria a data product met – i.e. how many “Yes” answers, and how many almost/sometimes met – i.e. how many “Partially” answers.
The ranking of data products by counts of “Yes” and “Partially” response indicates that the more established data products with national or global governance – such as OBIS, the National Reef Monitoring Network and Seamap Australia’s National Benthic Habitat Layer –received the highest number of “Yes” responses. Newer data products that are still evolving and seeking ongoing support - such as Squidle+ deployments and the three GlobalArchive data products - had more “Partial” responses. The AMP Ecosystems data product had the highest number of “No” responses.
By implementing the quality assessment procedure on the 20 priority data products and associated repositories, the project was able to identify 111 recommendations that if implemented would help make these data product “fit for purpose”. These recommendations, together with additional recommended changes to the Natural Values Common Language, improvements to the way in which: (i) canyons and seamounts, (ii) shallow coral reef habitats, and (iii) oceanic vegetation in the Coral Sea are identified in the AMP ecosystems data product - and the use of higher resolution bathymetry and multinomial habitat distribution models to map (with associated level of uncertainty) habitat forming species within this data product – are colloquially termed as the “Roadmap” for improving knowledge transfer to support Australian Marine Park decision making and management effectiveness evaluation (Table 1).
Table 1 Summary of the Roadmap issues identified during the course of NESP MaC Project 2.3
Issue
Summary
Status of the Natural Values Common Language
The Natural Values Common Language (NVCL) and its associated data products were designed to support a monitoring prioritisation process. Their status, and role within, the Authorisations, Incidence Response and State of Knowledge programmes has yet to be clarified or codified. At this juncture it is important for Parks Australia to confirm if it intends to proceed with its current classification of marine benthic habitats (via the NVCL ecosystems) or seek to adopt the more detailed (but data intensive) delineation of habitats provided by Seamap Australia via the National Benthic Habitat Layer.
Accuracy of the AMP ecosystems data product
The uncertainty associated with the ecosystems identified in the AMP ecosystems layer is unspecified. The NVCL implies that if an AMP ecosystem is shown as present within an area of interest, then all associated ecosystem components and sub-components are also present. A more rigorous approach is to use habitat distribution models to define the probability that habitat forming ecosystem components (such as seagrass, macroalgae and sessile invertebrates) are present with a specified level of certainty. The full use case for the AMP ecosystems needs to be articulated by Parks Australia to ensure that the utility of these models, and various representations of uncertainty that can be provided with them, will be adequate for future decision making.
Updating of the AMP ecosystems data product
A formal process to maintain and update the AMP ecosystems (including the addition of new ecosystems such as oceanic vegetation) need to be developed. This should include the process for inclusion of new data sources and the updating of any existing data sources. This process should be managed by Parks Australia.
Fit for purpose data products
The project identified 20 data products and repositories that Parks Australia rely on when addressing four of its six key questions. The quality assurance assessment performed in this project identified over 100 recommendations in order to make these products and repositories fit for Parks Australia’s purpose. As a general rule the older, more established and well supported data products and repositories attracted fewer recommendations than the newer ones.
Data accessibility
During the course of the project a number of potentially useful, but not publicly accessible, data products were identified. Private data products (such as these examples) can be categorised into four types: (i) data from science activities that were commissioned and funded by Parks Australia, DCCEEW, or NESP/NERP and conducted under NESP/NERP; (ii) data from science activities that were commissioned and funded by Parks Australia, DCCEEW or other third parties, but not conducted under NESP/NERP; (iii) data collected in AMPs by industry, commercial consultants or research institutions following authorisations issued by Parks Australia; and (iv) data collected under other regulatory regimes (such as fisheries management) that is relevant to AMPs. Parks Australia should consider strategies that would help make these types of data sets publicly available into the future.
In addition to individual data products and repositories, the project also examined the architecture and data product infrastructure that currently services Australia’s marine science and marine management community more generally. This analysis of the entire data supply chain, from requirements to end use cases, uncovered inconsistencies in the architecture used from minor issues to identifying data products and associated infrastructure that require significant work and support.
This broader review also identified eight very recent or currently in progress data infras
Funding
Commissioned by: NESP Marine and Coastal Hub
NESP MaC 2.3 Improving knowledge transfer to support Australian Marine Park decision making and management effectiveness evaluation : Department of Agriculture Water and the Environment
History
Confidential
- No
Commissioning body
NESP Marine and Coastal HubPagination
1-142:142Department/School
Ecology and Biodiversity, IMAS DirectoratePublisher
NESP Marine and Coastal HubPublication status
- Published online