ACADEMIA

Tracking environmental information about inland water is vital to manage and maintain water quality and ecology, but currently most of it is in scores of dispersed data 'islands'. "Some EU countries could have over 20 different databases, from computer records kept by a single researcher to massive data warehouses maintained by the Ministry of Environment," says Dr Palle Haastrup, a researcher at the Joint Research Centre's Institute for Environment and Sustainability, and coordinator of the IST program-funded Environmental Data Exchange Network for Inland Water (EDEN-IW) project. Most often, these different databases use a wide variety of software, in many languages and using different data formats. "Even the concepts governing specific terms vary. One country's measure of quality might relate to filtered water, while another country uses unfiltered." More prosaically, though very important, is simply getting access to the data in the first place. The EDEN-IW project sought to overcome these problems. The project developed special software agents that could take and 'translate' a user query so that all databases can be accessed simultaneously. The agents needed to cope with language, concepts and mixed data formats. To tackle the problem the project used open standards like XML, a language for defining the content of Web pages or even fields within a databases, and OWL, a protocol for developing ontologies, special vocabularies that allow machines to answer queries with context-relevant information. The project went even further, allowing users to analyse data sets, infer missing data and model various scenarios. It's an excellent tool to compare environmental metrics across national borders and to study how different countries respond to different problems. The ultimate vision was to provide a wide variety of tools for analysing inland water issues at a local, national and even pan-European level. It was an ambitious task and the EDEN-IW project achieved most of its aims, but it proved impossible to realise the overall vision with current technology. "In the laboratory we got the software working across a variety of different platforms, using different software in different languages, so we have a working prototype," says Dr Haastrup. What's more, the prototype developed by the project could also be applied to coastal waters research, general environment research such as air pollution, "… or any research or data mining that needs to access heterogeneous databases," says Dr Haastrup. But issues outside the scope of the project means the software cannot yet be released 'in the wild'. "Security is one issue, for example, and some of the partners will continue to work on that," says Dr Haastrup. Even so, very valuable lessons have been learned about what works and what doesn't, and the project will have an influence beyond it's central objective. "Our work informed the formulation of the European Water Framework (EWF), for example," says Dr Haastrup. The EWF will harmonise information across national boundaries and provide a common for comparing data. As such, it will provide a pan-European view of water quality issues.