Future priorities for research relating to
IST Environmental Applications
Held in Brussels on April 27th 2001
Recommendations for the preparation of
the 6th Framework Programme (2002-2006)
Workshop Report
EUROPEAN COMMISSIONINFORMATION SOCIETY DIRECTORATE GENERAL
Directorate B: Systems and Services for the Citizen
Applications relating to environmental protection
Intelligent Systems for Environmental Applications
Acknowledgements
On behalf of the unit B4, I would like to warmly thank all the participants in this workshop for their invaluable contribution in re-assessing the priorities for future research and development in the field of Environmental Applications with a clear focus on European added value. The response to our invitation to share ideas was exceptionally high, even many of those who were unable to attend the meeting insisted on contributing in writing.
This report is an attempt to summarise not only the discussions during the workshop but also to extract the main messages from the written contributions that were not discussed due to the lack of time.
I am greatly indebted to our “rapporteur” Prof. Keith Jeffery who kindly accepted to carry out this important task.
Guy Weets
Editor
This report and annexes are available on the Internet
at the following address: http://www.cordis.lu/ist/ka1/environment/home.html
Intelligent Systems for Environmental Applications
Workshop on Intelligent Systems for Environmental Applications
Workshop Report Keith G. Jeffery
______
Executive Summary
A workshop held in April 2001 was tasked with identifying the priority actions for the next 5-10 years and identifying lines of action for the workprogramme 2002 classified as RTD Actions, Support Measures and Thematic Networks.
The rationale for action is based on perceived areas where Europe has gaps in the IT armoury for managing environmental applications such as data collection and sensors, provision of data not information, the problems around public-private organisation partnerships, the multiple and incompatible standards, the problem of a lack of usable interoperable tools and the need for a ‘one stop shop’ for access to the data, information, tools and facilities required.
The gaps were analysed to determine where investment might best be utilised. The areas for investment identified were: advanced sensors, improved and commonly accepted data standards, the tools to provide from data information, metadata associated with the data and information to expedite utilisation, software tools an systems which interoperate with data standards and each other, an overall accepted architecture based on open interfaces and standards which would allow the development of integrated systems. Some political obstacles were also identified and need addressing, they concern mainly the need for international co-operation because the environment does not respect national boundaries.
The proposed themes to be supported are:
(a) Environment open architecture, providing the basic interfaces and standards;
(b) Environment data warehouse, providing the unified data model for a virtual warehouse across heterogeneous distributed data stores;
(c) Environment knowledge base, providing decision makers and managers with a support environment of knowledge generated reliably from data and information;
(d) Environmental workbench for tool development, permitting the rapid development of a useful toolkit for environmental information and knowledge management, analysis, display and utilisation;
(e) Environment data acquisition, providing improved, intelligent, configurable sensors and sensor arrays for improved data quality and relevance
The major recommendations for implementation were:
(1) a few, large-scale integrated RTD projects;
(2) accompanying measures especially for dissemination of best practice;
(3) thematic networks of excellence to bind the communities and to ensure maximal discussion of and ‘buy-in’ to standards;
(4) a few small, high-risk, innovative RTD projects to explore novel technologies for environmental applications.
Report
Objectives of the workshop
The objectives of the workshop were introduced as follows:
(1) Identify priority actions for the next five to ten years
with a clear EU dimension and added value and taking into account the new instruments (or methods of execution) available such as integrated projects, networks of excellence and support of national programmes (article 169);
(2) Identify lines of action for the work-programme 2002
classified into RTD actions, Support measures and Thematic networks.
The workshop was structured around three overlapping themes:
Theme 1: representation and communication of environmental information and knowledge;
Theme 2: data acquisition and pre-processing, sensors, telematics infrastructure;
Theme 3: generic applications making use of the above.
There were excellent prepared contributions sent and collated beforehand and stimulating discussion at the workshop from which clear proposals emerged.
Rationale for action
The need for action is based on several factors. There are global imperatives in managing the environment which require understanding of environmental processes. This in turn requires scientific observation followed by hypothesis creation, testing by observation or experiment (including simulation) then utilising the knowledge so gained for benefit – either commercial or for the good of the European Society.
The need is not only technical / scientific. There is a requirement for political action to agree common standards (which raises many issues concerning variations in European countries including multilinguality) for data access and interchange and to reconcile the issues of rights to use with rights of the data owner. There are additionally problems concerning policies of individual countries with regard to accessibility of data in the public - commercial interface.
Behind the need is the basic point that Europe, with a few notable successful exceptions, has not generally been leading in using environmental data for commercial purposes in managing the environment - with implications for cost-reduction or increased profitability. Europe has not led in the connection of environmental information directly to decision making in, for example, the agricultural industry, in fishing, in transportation or the leisure industries.
Data Collection and Sensors
Europe has a rather good record in collecting and storing environmental data. Well-known data centres of archival material exist. There have been examples of utilisation of novel sensor and associated communications technologies but in general there has not been the take-up of modern sensor technology that one might expect. Modern sensors have built-in logic to allow better information on calibration, accuracy and similar factors to be recorded so improving the utility of the data. Some have self-correcting facilities, and some accept downloads of intelligence from computer networks should a central validation service discover errors or instrument drift.
Data not Information
Europe has been generally lagging behind others in adding value to environmental data by converting the date to information (structured data in context) which makes it useful for many applications and commercial opportunities. Associated with this conversion of data to information is the assurance of (or value of) quality and the availability of provenance and other associated information to assist the end-user in utilising the base data.
The Partnership Problem
The result of the data provision being inadequate for business decision-making has led to the ‘partnership gap’ where commerce and industry (and not-for profit organisations) fail to invest in converting the data to information for their own purposes simply because of cost and effort to do so. Similarly the agencies producing the data – often publicly-funded - have no incentive to convert the data to canonical information ready for easy use for decision-making by a wide range of commercial and non-commercial end-user organisations.
The Standards Problem
Standards do exist for environmental data (and associated metadata) and are used. The problem is there are many different standards, even within one discipline, despite attempts to co-operate and agree on one common standard in some areas. Without standards it is impossible to create commercially-viable toolsets to assist in manipulating environmental information in a way suitable for decision-making or other commercial activity, or for non-commercial research and development or environmental protection. Within Europe there are particular problems in standardisation of co-ordinate data (spatial and temporal) as well as the multilingual (leading to syntactic differences) and multicultural (leading to semantic differences) problems.
The Tool Problem
There are many different software / system architectures used in environmental science with little or no agreement on tool functionality or interfaces. Standardising metadata and datasets would provide a way of formalising at least one interface set, and the user interface to the tools is almost certainly to be web-based. Thus the remaining requirement is for the functionality of the tools to be agreed and defined.
The One-Stop-Shop Problem
For many consumers / users of environmental information it would be convenient to have one portal (or at most a few portals e.g. subject area) to access all the information and associated data processing needed to solve any environmental problem. This implies a unifying architecture to hide from the end-user the heterogeneity of datasets and processing tools and also the complexities of rights of access.
Assessment of current achievements and gaps
A Technical
1. Sensors: although sensor development is quite sophisticated there is much to be done in the provision of low maintenance, self-correcting / calibrating sensors with logic to allow choice of detection rates, ranges, precisions. The need for internet-connected sensors is paramount, probably with sensor arrays feeding an internet-connected server. The control of sensors by end-users for experimental purposes is part of e-Science.
2. Data Standards: the data available will be utilised less than optimally if there are not well-defined and agreed data standards to allow interchange between data stores and also to allow various software tools to act on the same dataset. At present there are multiple proprietary standards so providing the potential user with a discouraging high effort threshold.
3. Information: the structuring of data into information (‘data in context’) is critical for the utilisation together of datasets from different sensor types or from different environmental disciplines. The design of experiments and sensors requires as a background a knowledge of formal data modelling to ensure usability.
4. Metadata: metadata (‘data about data’) atached to datasets allows diferent datasets to be used together and with appropriate software tools in an intelligent way mediated by a person or by a software system. Some metadta schemes exist in Europe (and globally) but the metadata standards are usually more designed to be computer-readable and human-understandable than computer-understandable (and thus automatable).
5. Software Tools and Systems: at present there are many disparate software tools which work only with datasets of a certain form and kind. There is a great need for the development of software tools within systems where interfaces between software tools and datasets are standardised and / or mediated by metadata. Tools are required for: Validation, Analysis, Display, Simulation, Data mining, Prediction, Decision Support, Virtual reality and Pervasive mobile support
6. Overall architecture with defined data and software interfaces: there is a need or an overall architecture to be proposed and agreed encompassing much of the above
7. Integrated Systems with interfaces for plug and play: given a unifying architecture and standard interfaces for datasets and software tools then ‘plug and play’ becomes possible. This allows, for example, an end-user to utilise a range of analysis tools (each with its own algorithm) against a given dataset (or group of datasets derived by other software tools) applying them in turn to find the best results. This is not possible at present.
B Political
The key political gaps are as follows:
a)Recognition that environmental systems do not respect national boundaries: rivers cross frontiers as do air masses. Management of the environment is not simply within control of a nation-state. As information becomes progressively more easily available (by the measures outlined above in section 3.1) it will be impossible for any nation state to prohibit its citizens knowing about the environmental state of features (e.g. rivers, coastline, air quality) which traverse their state and adjacent states.
b)Need European-wide standard systems for co-ordinates (space, time): it is important for co-ordination of detection and for linking models, it is critically important for environmental decision support systems, that a unified scheme for co-ordinates is used in Europe. Right now there are different geographical projections in use (and the transformations tend to be approximate) and there are multiple time zones and different precision of time-recording.
c)Need European standards for data types for environment: in order to ensure data quality is understood and in order to allow people and computers to understand whether the same term means the same thing in different systems (with appropriate multilingual facilities) there is a requirement for the development of an environmental thesaurus, and an environmental domain ontology.
Public/private organisation interaction/co-operation: environmental data and associated processes are measured, influenced, controlled and understood by a variety of organisations. Currently there are some barriers between different publicly-funded bodies and also between different commercial bodies. Similarly there are barriers between public bodies and commercial bodies. There is a need for a clear framework for the ‘rules of engagement’ when such bodies work on environmental data and processes.
Priorities for future research
From the written contributions prior to the workshop mediated by discussions at the workshop it is possible to extract a few themes which should be considered as priorities for future research - with a view to benefiting not only the European wealth-creation but also improving the environment for the citizen, including the availability and ease-of-use of environmental information.
Environment Open Architecture
Leverage existing national efforts - publicly-funded or commercial - by developing an open architecture and supporting tools to help federating existing or up-coming national environmental information systems. This implies that it is searchable by means of semantically meaningful metadata: (e.g. an EML dialect of XML) and offers sophisticated methods for auto-classification (Topic maps, data mining to generate ontologies). Such an architecture would be building on existing international standards.
Such a theme is justified by the relatively small investment required to leverage pre-existing and committed investments by nation states in order to create a better environmental systems architecture for all to use. It would foster better trans-Europe co-operation in all aspects of environmental monitoring, modelling, decision-making and management, it would encourage commercial development of software and also the development of even better interfaces and standards.