1. Kingsley Dunham Centre Keyworth Nottingham NG12 5GG Tel 0115 936 3100 © NERC All rights reserved Developing a generic Geoscience Markup Language: GeoSciML Tim Duffy, Marcus Sen, James Passmore British Geological Survey trd@bgs.ac.uk,mase@bgs.ac.uk,jpass@bgs.ac.uk

2. © NERC All rights reserved BGS Introduction The British Geological Survey (BGS) has been collecting geoscience data since it was founded in 1835, and is the UK national centre for earth sciences covering all the major disciplines within geology and geophysics. It operates both in the UK and internationally. A major role of the BGS is one of data custodian, and in recent years it has put major effort into making this data more accessible both to internal and external users. A key component to this is the use of web technologies including XML for data transfer.

3. © NERC All rights reserved How we handle corporate data Corporate relational database (Oracle) Attributed map data using CAD & GIS (Intergraph and ESRI) Individual projects use a whole host of heterogeneous formats from Excel to custom program formats etc. Ref: http://www.bgs.ac.uk/magazine/magazines/docs/ewise16.pdf http://www.bgs.ac.uk/magazine/magazines/docs/ewise16.pdf

4. © NERC All rights reserved XML experience Specific project use e.g.: Textbase – storing and retrieving report and other text attributed with subject metadata and linkages to corporate databases DEAL – exchange of North Sea oil and gas well information for UK Department of Trade and Industry and oil companies XMML consortium sponsors http://xmml.arrc.csiro.au/

5. © NERC All rights reserved XML for data exchange File format Using XML means file parsing can be carried out by common tools eliminating part of the work involved in data exchange Human readability and partially self-documenting nature of XML documents makes it easier to read someone else’s data and makes it a good candidate for archival use Data model Common tools (e.g. DTDs, XML Schema) can ensure data conforms at least partially to a particular data model Still can’t exchange data between incompatible models but common tools (e.g. XSLT) exist to make translation process easier if a mapping or partial mapping between models exists But having a standard intermediate model or library of model pieces would reduce the number of translations that need to be coded

6. © NERC All rights reserved GML Geography Mark-up Language (GML) is an OpenGIS Consortium (OGC) standard using W3C XML Schema Create your own application schema using components from GML eXploration and Mining Mark-up Language (XMML) is a GML application with a focussed geoscience scope GeoSciML is also being developed as a GML application with a wider generic geoscience scope that imports and builds on XMML and GML

7. © NERC All rights reserved GML components Features Geometry Coordinate reference systems Topology Temporal information Definitions and dictionaries Units, measures, values and directions Observations (also separate Observations & Measurements standard) Coverages Styling

8. © NERC All rights reserved XMML Components Customisation of existing GML components (features, observations, temporal components,...) Geologically useful geometric constructs from GML components (surfaces etc.) Features, values etc for subject areas such as geochemistry, earth material description, mineral exploration projects, geophysics etc.

9. © NERC All rights reserved Feature model Developed by ISO TC/211 (ISO 19109) Most suitable for discrete objects in space Compare traditional CAD & GIS “geometry-first” approach

10. © NERC All rights reserved A GeoSciML Schema.........

11. © NERC All rights reserved GeoSciML Development Approach Select a few trial geoscience subject areas Boreholes Structural geology Text Dictionaries Geochemistry, geophysics, geotechnical... Use existing software to generate prototype GeoSciML instances from existing data (N.B. currently we don’t have software aware of XML Schema type hierarchies and substitution groups.) XSLT, Oracle XSQL, ArcIMS OGC WFS servlet

12. © NERC All rights reserved Boreholes We (unsurprisingly) have quite a few borehole related data from simple index information on (almost) all boreholes drilled in the UK to geological log, water level, geochemical etc. data from boreholes Currently delivered by a variety of front-ends: web-forms leading to HTML tables, graphical views (illustrated), Internet GIS (ArcIMS based)...

13. © NERC All rights reserved Boreholes Instance missing None NS88SE/SE/6792/ 3/2 FALKIRK,BANKSIDE IND ESTATE. 289495 0681545 7.94 SEE ALSO SE 6793. 6...

14. © NERC All rights reserved Text BGS has developed an in- house application (Textbase) to retrieve report fragments and other text based on subject meta-data and what features it describes OGC trialled an application called Location Organizer Folders (http://www.opengis.org/docs/0 1-037.pdf) to bring together assorted media types in a GML frameworkhttp://www.opengis.org/docs/0 1-037.pdf

15. © NERC All rights reserved Text Instance 1957 3601 LI MMG SSG Structure contour maps of the top surface of the Sherwood Sandstone Group, Mercia Mudstone Group and the Lias (Figures 7c-e) show gentle (1 to 2°) easterly dipping surfaces ‘ rippled ’ by numerous open folds, commonly associated with minor, dominantly normal faults, orientated north-east or north-west, with displacements typically less than 20 m....

16. © NERC All rights reserved Structural Geology We are starting with the model (GSD2) currently being implemented for BGS’ field mapping data

17. © NERC All rights reserved Structural Geology Instance 443554.612 1179784.018 466908.386 1217077.305 NMMUS SCHH 455313.386,1195067.109 455314.533,1195067.697 455316.358,1195064.138 455313.386,1195067.109...

18. © NERC All rights reserved Dictionaries Currently just simple definition lists or hierarchical structure but not complex relationships between defined terms (thesaurus)

19. © NERC All rights reserved Dictionaries Instance <gml:DefinitionCollection gml:id="BGSRCSSubset-XFDAE-ADFEE" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:gml="http://www.opengis.net/gml"> A subset of entries from the BGS Rock Classification Scheme BGS Rock Classification Scheme CLAY (UNDIFFERENTIATED) CLAY CLAY, GRAVELLY CLYGV...

20. © NERC All rights reserved Other areas We are also interested in geochemical, geophysical and geotechnical data. There is currently active development by XMML project of geochemical and geophysical schemas which look likely to cover all our needs. Geotechnical data may be based on UK body Association of Geotechnical and Geoenvironmental Specialists (AGS) developing XML format but needs work

21. © NERC All rights reserved Exchange process gml:location 320000,179000 520000,345000

22. © NERC All rights reserved Conclusions Designing and understanding a comprehensive geoscience data exchange model is difficult Need to drive from use-cases and prototype exchange formats for geoscientists to try out with their particular use-cases We propose that a similar development process be extended to the wider geoscience community, with the support of bodies such as the International Union of Geological Sciences' (IUGS) Commission for the Management and Application of Geoscience Information and the OGC, so that GeoSciML can become a common language for generic geoscience information exchange using XML based web processes.

23. © NERC All rights reserved