STUDY ON INTERIM EVALUATION OF EUROPEAN MARINE OBSERVATION AND DATA NETWORK - final report

full report (pdf)

In order to improve access to marine environmental data, the European Commission has proposed the establishment of a European Marine Observation and Data Network (EMODNET) that will ensure that marine environmental data is compiled in a comprehensive and compatible system, and made accessible as a tool for better governance, expansion of value-added services and sustainable development.

In support of the preparatory actions of the EU’s Integrated Maritime Policy a set of projects were launched in 2008 and 2009 with the aim of creating five portals that assembled fragmented and inaccessible marine data into interoperable, contiguous and publicly available data streams for complete maritime basins.

Each of the five portals is concerned with a different aspect of marine environmental data: chemistry, hydrography, biology, geology and physical habitats. Each portal is operated by a different thematic group. The thematic groups have been tasked by the European Commission with gathering relevant marine measurements within at least two sea–basins, and making the data available through a portal enabling public access and viewing of these data.

The aim of this study was to conduct an interim evaluation of the progress the portals had made to fulfill the objective of enabling public access to the marine environmental data of the portals. The evaluation was given three tasks to perform on each of the portals, these were:

• Task One to assess the user-friendliness of the portals;
• Task Two to gauge the re-usability of the data; and
• Task Three to determine how well the portals have overcome the legal obstacles to data sharing identified in the study “Legal Aspects of Marine Environmental Data” (FISH/2006/09).

As Task One required some subjective assessment of the user friendliness of the portals, the study also contacted potential users of each of the portals and asked them to complete a short questionnaire on the portal. The results of these have been included in the study. However, the low level of responses does not allow a quantitative assessment but does provide an additional user perspective.

Task One

All of the portals except the physical habitats portal have an EMODNET-related url and homepage that immediately identifies them as part of EMODNET. The physical habitats portal is under the domain of its host organisation the JNCC. The biology, chemistry, and hydrography homepages have clear navigation to different sections of the portal, where as the geology homepage is just a single page with a link to the OneGeology Europe portal. Once in the OneGeology portal, it is not clear what the relationship to EMODNET is, or even were the EMODNET data is located.

The information of data quality provided on the portals was generally clear for the data products on the portals. The information on data quality for the data sets was sometimes harder to interpret, as the data set may have come from several organisations. Where this was the case providing the guidelines that the site portal used for quality control and assurance was useful, as was done by the hydrology and chemistry portals.

Task Two

Task Two was to gauge the reusability of the data, and this was done by trying to perform predefined tasks. The types of task differed for each portal; an example for the hydrology portal was to download the bathymetry data for a sea basin and plot it a GIS application. For each task a record was kept of the steps required to perform the task using the portal and the data it provided. As each of the tasks were different, it was not possible to make direct comparisons between the portals. However, they are provided here in the order of ease with which it was possible to complete the task.

The Task for the biology portal was simple to achieve, and could be completed using just the functionality of the portal itself. To test the portal further, an additional task was carried out show the presence absence of selected bird species for and area. The portal again made this task simple and it was possible to achieve in a few steps in a GIS application with simple GIS functions.

The Geology portal only had one map of sediments for the North and Celtic sea areas. It was simple to download this map and clip it so it just showed the sediments for the North Sea.

The task for the hydrology portal was to create a bathymetric map of a sea basin. This task was completed using only the EMODNET Gridded Bathymetry data product one of the portal. The files were downloaded for one sea basin and imported into a database application to produce one complete dataset for the sea basin. The data was then imported from the database into a GIS, and, using one of the more advanced futures of the GIS application, a raster layer depicting depth the data was created.

The physical habitats portal uses some of the data products from the other portals for the creation of the modelled seabed habitats that are the principal data product of the site. The maps are downloadable as zip files that contain shape files of the habitat data. The files are easy to open or import into a GIS application. To complete the task a few steps had to be completed, such as projecting the shape file into a suitable map projects, before running a simple analysis function to perform calculations on the data in the files.

The task for the chemistry portal was to create some spatial and time plots of pollutant concentrations in the water column and in sediments. The majority of data on the portal is in data sets at single points, of which there are currently 468,360. The search functions of the portal are advanced and have features that enable searches to be refined and saved that help to indentify suitable record sets. However, it was often only when the data sets were downloaded and opened that their applicability to the task could be gauged. Once suitable data sets had been identified, the processing of the data files was assisted (although not required) by using the Ocean Data View software which itself has good tools for display the data, or it can be exported to spreadsheet or GIS application to display the data.

Task Three

As described in an earlier Study, ‘Legal Aspects of Marine Environmental Data’, data of the kind contained in the portals are subject to intellectual property rights (IPR). The owners of these IPR, who will often be persons other than the members of the consortia, are basically free to determine whether or not the data may be used or re-used by a third party in accordance wit their own data policies and the contents of contractual licence agreements that these call for. The basic objective of Task Three was to assess how the approach of the portals to the IPR of the owners of the data to which they provide access in terms of the use and re-use of those data.

Analysis of the portals, as well as the telephone interviews with representatives from the consortia responsible for the development of the portals, shows IPR issues have clearly been taken into account in terms of the design and operation of the portals. The precise approach to IPR varies from portal to portal just as the layout of the portals and the mechanisms for accessing to the data does ranging from the rather sophisticated ‘shopping cart’ systems on the chemistry and hydrography portals, with a prior registration procedure, to simpler ‘click licences’ on for example the physical habitats portal.

Conclusions

The portals have been set up by different consortia of research organisations and consequently, with the exception of the chemistry and hydrography portals, they do not share a common approach. Many of them, such as the biology portal and the geology portal, were or are based on existing projects.

For tasks 1 and 2 all of the portals were found to be intuitive to use and where help was needed it was usually available within the site. Requests for help when using the portals to complete specific tasks received quick and useful responses. Where problems did arise it was often down to the quality or completeness of the metadata data rather than the technical function of the portals themselves. Most of the tasks were straightforward to achieve and tools and data products within the portals made completing the tasks much easier they would of otherwise have been.

The chemistry and hydrography portals had the greatest functionality, with several different methods of searching for data that were intuitive and simple to use. The searches could also be refined to filter results from initial searches, although this functionally was less intuitive than other features on the site. Searches could also be saved: this was found to be a useful feature in providing additional references for the data used for Task Two.

The shopping basket system also made managing the data that was requested and downloaded simple. When making the initial request through the system it was immediately clear whether or not the data sets were subject to any restrictions, and at this point is was possible to refine the request to, for example, focus on unrestricted data sets. It was easy to follow the progress of requests and the maintenance of a history of requests also helped managing the data that had been downloaded from the portals.

The chemistry, hydrography and biology portals all provided access to both data sets and data products. The physical habitats and geology portals only had data products. The restrictions on most of the data sets in the hydrography portal meant that data products were the principal resource used on that site.

It was moreover possible to retrieve datasets from the chemistry and biology portals. The datasets in the biology portal are presented as either complete datasets or amalgamated by taxa, meaning that fewer steps were required to use the data. However, when completing the tasks required for this study inconstancies in the data were found (e.g. missing dates in the birds data) which may reduce the usefulness of the data overall. However, overall less than 10% of records have missing date records.

The chemistry datasets are closer to the original data sets, meaning that a lot of detail is potentially available. As the data in the dataset is generally quite complex it and can only be viewed data once the file has been downloaded, it may take several attempts to obtain the data required. However when the correct dataset is identified it is more likely to be comprehensive, and may be more useful as a result.

Because of the differences between the portals, and the fact that they do not share a common approach, it is hard to make direct comparisons. Nevertheless we have summarised our findings in a comparative matrix in Table 16.

Table 1 Matrix comparing portals

As regards Task Three main the main issues are: (a) the apparent lack of uniformity in the approach of IPR and use/re-use policies of the various portals as "data publishers"; and (b) the apparent lack of clarity on the identity of the licensor. Finally one other issue that the portals will need to address is the collection and processing of personal data of users.

Recommendations

Recommendations - biology portal

• Provide information on what terms in metadata mean (e.g. ‘status’).
• Review metadata and use IHO sea areas (or other) to standardise geographical areas.
• Review metadata to make temporal ranges more searchable.
• Review terminology – several common names for one species etc.
• Review thematic areas – macroalgae/algae etc.
• Make information on quality checking of data available on the website.
• Introduce a uniform, simplified and comprehensive set of user terms instead of the current approach of linking to the Eurobis and MarBEF terms (for raw data) and to Emodnet.eu (for online data products).
• In terms of licensing arrangements, clarify the identity of the licensor/licensors.
• Address the issue of the collection and processing of personal data.

Recommendations – chemistry portal

• Include information on how quality-checking is undertaken.
• In terms of the licensing arrangements, clarify the identity of the licensor/licensors.
• Clarify and align the provisions of the data policy and the licence agreement into one uniform and comprehensive set of terms and conditions.
• Address the issue of the collection and processing of personal data.

Recommendations - hydrography portal

• Break up long pages of text with headings or use section breaks and links at top (e.g. ‘EMODNET’ page)
• In terms of the licensing arrangements, clarify the identity of the licensor/licensors.
• Clarify and align the provisions of the data policy and the licence agreement into one uniform and comprehensive set of terms and conditions.
• Address the issue of the collection and processing of personal data.

Recommendations - geology portal

• Make marine data more prominent on website (or directions on how to get to it) as users may not know that EMODNET refers to marine data.
• Provide a link to the DG MARE EMODNET web site pages.
• Adding guidance/help on how to use the portal would be useful.
• Once data becomes available, include metadata in same portal as terrestrial geological data and make it clear how to search catalogue to just retrieve the marine datasets, also linkable from the portal homepage (with EMODNET banner) rather than just 1GE.
• In terms of the licensing arrangements, clarify the identity of the licensor/licensors.
• Clarify and align the provisions of the data policy and the licence agreement into one uniform and comprehensive set of terms and conditions.
• Address the issue of the collection and processing of personal data.

Recommendations – physical habitats

• A Separate searchable data catalogue would give the site the same general functions of all the portals
• Main page under its own EMODNET URL and uses similar layout to other EMODNET portals
• In terms of the licensing arrangements, clarify the identity of the licensor/licensors.
• Clarify and align the provisions of the data policy and the licence agreement into one uniform and comprehensive set of terms and conditions.