FOR ONTOLOGY-BASED CLASSIFICATION , INTEGRATION , AND INTERDISCIPLINARY USAGE OF GEOSCIENCE METADATA

The Semantic Web is a W3C approach that integrates the different sources of semantics within documents and services using ontology-based techniques. The main objective of this approach in the geoscience domain is the improvement of understanding, integration, and usage of Earth and space science related web content in terms of data, information, and knowledge for machines and people. The modeling and representation of semantic attributes and relations within and among documents can be realized by human readable concept maps and machine readable OWL documents. The objectives for the usage of the Semantic Web approach in the GFZ data center ISDC project are the design of an extended classification of metadata documents for product types related to instruments, platforms, and projects as well as the integration of different types of metadata related to data product providers, users, and data centers. Sources of content and semantics for the description of Earth and space science product types and related classes are standardized metadata documents (e.g., DIF documents), publications, grey literature, and Web pages. Other sources are information provided by users, such as tagging data and social navigation information. The integration of controlled vocabularies as well as folksonomies plays an important role in the design of well formed ontologies.


INTRODUCTION
The Electronic Geophysical Year (eGY) 2007-2008 (http://www.egy.org/declaration.php),an initiative of the International Union of Geodesy and Geophysics, was founded to strengthen the awareness of a sustainable handling of geoscience data, information, and knowledge and the important role of Earth and Space Science Informatics (ESSI).ESSI connects informatics research, development, and practice with the processing and solution of very specific, but also general, science and societal questions related to the research areas of the Earth and space domains 50 years after the International Geophysical Year 1957-1958.My experiences as European eGY representative and as organizer and participant of various eGY and ESSI related workshops and congresses within the last years, along with the work of the authors at the German Research Center for Geosciences' data center (www.gfz-potsdam.de) in the design, implementation, and operation of the GFZ Information System and Data Center for Geoscience Data and Information, show clearly the demand for an intensified engagement for the creation of conditions for a sustainable handling of the life cycle of geoscience and space science data from the experiment, through the processing of data and information, to the generation of knowledge.Such ESSI objectives as the classification of geoscience data and information, the metadata extension of data, the development of concept models for the discovery of relations between data classes, and the usage of taxonomies and ontologies play an important role for an on-demand discovery, a sustainable management, and a systematic use of multi-domain data and information for the generation of knowledge for a better understanding of the complex system Earth.In particular, the intense study and practical implementation of the scientific and technical ideas of the Semantic Web approach, according to Daconta et al. (2003), are very important for the creation of interoperable and inter-domain geosciences data, information, and finally, knowledge.This paper demonstrates the usage of metadata classes and related standards according to a Semantic Web approach for the description, classification, and design of the conceptual structure of the geodetic and meteorological data and information of the Global Geodetic Project (GGP) in the framework of the International Association of Geodesy (IAG).

IAG, GGOS, AND GGP PROJECT
The International Association of Geodesy (IAG) (http://www.iugg.org/associations/iag.php) is the scientific organization for the exchange of scientific results and research ideas as well as the promotion of scientific work in geodesy and related geoscientific fields.Geodesy studies the geometry, the gravity, and the rotation of the Earth.Rothacher (2002) calls these three research areas the three pillars of geodesy.In particular, the processing and supply of a highly accurate terrestrial reference frame (TRF), derived by combining results from the three geodetic pillars, are very important for the detailed description and presentation of spatial terrestrial data and information.Computation of the TRF is the IAG's contribution to the creation of the Global Earth Observing System of Systems (GEOSS) (www.earthobservations.org).The IAG consists of commissions, scientific working groups, a central bureau, and services.The IAG flagship project is the design and implementation of a Global Geodetic Observing System (GGOS).The GGOS integrates geodetic techniques, models, and approaches for better long-term consistency, reliability, and understanding of geodetic, geodynamic, and global variations of the Earth.GGOS provides the scientific and infrastructure basis for global change research in Earth sciences.The main objective of the GGOS project, related to service and infrastructure aspects, is the promotion of geodetic data products and appropriate IAG services.Project requirements are the implementation of sustainable services for the collection, validation, archiving, and distribution of geodetic observation data, processed data products, such as global gravity models, and finally cross-domain knowledge products, such as results according to temporal variations of ocean currents for the estimation of the impact for climate changes (http://www-app2.gfz-potsdam.de/pb1/op/grace/).

Figure 1. IAG Data Services Structure
Figure 1 shows the current structure of IAG services in a GGOS environment, which are planned also to become services at GEOSS.Geometric related services are: • International Earth Rotation and Reference Systems Service (IERS), • International GNSS Service (IGS), • International Laser Ranging Service (ILRS), • International DORIS Service (IDS).
Data Science Journal, Volume 11, 18 October 2012 Gravity related services are: • International Gravity Field Service (IGFS), • International Gravimetric Bureau (BGI), • International Geoid Service (IGeS), • International Center for Earth Tides (ICET), • International Centre for Global Earth Models (ICGEM).Sea Level related services are: • Permanent Service for Mean Sea Level (PSMSL), • International Altimeter Service (IAS).Other service elements are: • Time Section of the International Bureau of Weights and Measures (BIPM), • IAG Bibliographic Service (IBS).
The International Center for Earth Tides (ICET), as part of the gravity service area of GGOS (Figure 1), is also a World Data Center service, known as WDCC for Earth Tides (http://www.astro.oma.be/ICET/).Furthermore, the results of the international Global Geodynamic Project (GGP), an IAG Inter-Commission project of "Commission 2: The Gravity Field" and "Sub-Commission 3.1: Earth Rotation and Earth Tides", provide the scientific basis for the calculation of Earth tides.The main objective of the GGP project (http://www.eas.slu.edu/GGP/ggphome.html) is to record the changes of the gravity field of the Earth with a number of highly accurate superconducting gravimeter (SG) observatories (Figure 2) distributed worldwide.
The data are also used for researching global motions of the whole Earth, such as the Chandler wobble (named after an American astronomer who in 1884-1885 discovered a movement in the Earth's axis of rotation that causes latitude to vary with a period of 14 months), up to the surficial gravity effects of atmospheric pressure and groundwater.The SG observatories are run independently by national groups of scientists and technicians who send gravity and meteorological data each month to the GGP Information System and Data Center GGP ISDC (http://ggp.gfz-potsdam.de)at the GFZ Potsdam in Germany.The provided gravity data differ in sampling rate and treatment after measurement.At the GGP ISDC there are calibrated raw data, corrected data related to jumps and gaps, as well as data which are processed according to an ICET standard procedure.These standard data files are processed at ICET (http://www.upf.pf/ICET/).In 2009, ICET operations were transferred from Brussels to Tahiti, the new location for ICET (http://www.icsu-fags.org/ps04icet.htm).All GGP data are public after a certain period of time (normally half a year) and can be requested and downloaded both at the specific GGP ISDC and the GGP section of the GFZ Information System and Data Center (ISDC) portal (http://isdc.gfz-potsdam.de/ggp)at the Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences.The basic architecture, the data philosophy, and the operation of the ISDC portal are described in the next section.

ISDC PORTAL
The German Research Center for Geosciences' Information System and Data Center (ISDC) portal is the door to GFZ's geodetic and geophysical data products and information from satellite missions and in-situ measurements, as described in Ritschel et al. (2007Ritschel et al. ( , 2008)).The GFZ ISDC portal integrates prior-independent and large-scale project driven ISDCs, such as the satellite mission related ISDCs for CHAMP (http://www-app2.gfz-potsdam.de/pb1/op/champ)and GRACE (http://www-app2.gfz-potsdam.de/pb1/op/grace)as described in Ritschel et al. (2006) as well as the already mentioned international project GGP.The ISDC portal consists of the portal framework, a content management system, user and product management, and presentation applications as well as a detailed monitoring and statistics software package.The ISDC portal Graphical User Interface (GUI) provides registration and authorization functionality on the top, navigation and monitoring elements left-hand and right-hand as well as a main content part in the center of the Web page (Figure 3).The portal framework is based on the open-source software PostNuke (http://postnuke.com/),now known as Zikula.Besides the portal framework and standard components (such as user registration and authorization, content management, and a user forum), most of the data related portal components (such as a retrieval and catalog system, product browser, product ordering system, monitoring, and administration) were developed by the ISDC team.A map server component (Figure 2), which is part of the retrieval system, provides spatial search and spatial presentation functionality.In addition to the portal frontend functions, the complete data management backend component including the row data archive components (long-term archive and online archive) is an in-house development.international scientific processing groups.More than 1900 users and user groups of the international scientific community are currently registered at the ISDC portal.Along with Germany, the greatest number of users are from China, the U.S.A, India, Japan, Canada, the United Kingdom, France, and Italy.Most important for cost-saving and efficient management of the very different product types and data products (for a detailed description see sub-section 5.1) were the development and usage of standards for content and semantic related metadata of product types and data products as well as for the ISDC backend data handling processes.This article does not go into details of the data handling process.Instead, the next section describes the ISDC product philosophy, which defines the standards for ISDC product types and data products.Figure 4 schematically illustrates the creation of an ISDC data product on the left-hand side and the processing of metadata and data products on the right-hand side.An ISDC data product always consists of one data file or a collection of data files and an appropriate metadata file.The metadata file structure is standardized and is partly based on the Directory Interchange Format (DIF) as described in the "DIF Writer's Guide" of the GCMD (http://gcmd.nasa.gov/User/difguide/difman.html).In an extension of the standard for ISDC data products, all product types are described by DIF metadata documents.The sources of metadata, both for data products and product types, are mainly data file headers, metadata in processing software, non-standardized metadata files, and even unstructured information often stored only in scientists' heads, as shown on the left side of Figure 4.
Because of the introduction of standardized data product metadata files, the ISDC backend software only needs to operate with these standardized metadata documents in order to control the data flows and to load the ISDC catalog (metadata database) and ISDC archive systems.Finally, the ISDC data product catalog provides ISDC data product retrieval and request functionality for its users in order to search and download data and metadata files (right side of Figure 4).The ISDC product type metadata documents contain the semantic description of the product types.These metadata documents are XML files and match the DIF standard (Mende et al., 2008).In contrast, the data product metadata files of data products related to the new ISDC collaboration projects, such as orbits of the Terra-SAR-X satellite and Earth rotation parameter, contain only semantic information of the data files, supplemented with compulsory metadata values for the independent viability of these documents.

DATA LIFE CYCLE
Data is produced during scientific experiments or simulations.The output of the scientific processing of (raw) data is data and information, such as domain-related values, profiles, and models.These data and information are used in order to create domain-related but also cross-domain scientific knowledge, often published in books and journals, as well as societal knowledge, used, for example, for decision support activities.More and more data, information, and knowledge are produced every day.According to Gantz et al. (2008), the generation of digital data is increasing at an exponential rate.However, are these exponentially increasing data, information, and knowledge really available for usage and re-usage by users and applications?At least in the present, the answer is no.Identified problems are: • exponentially increasing number of data (files), • much data and information are isolated and not stored in a sustainable form, • much data, information, and knowledge are not publically available, • most data, information, and knowledge are presented in proprietary structures and systems only, • most data, information, and knowledge are incompatible and non-interoperable, • it is often easier for scientists to get funding for new projects than for data re-using activities.5), according to Lyopn (2007), not only illustrates the life cycle of data from an idea, through the generation of data, to the transformation from data into knowledge but also addresses the data related processes, such as project elaboration and application and modeling.Artifacts of these data related processes are digital sources of semantics for cross-domain descriptions, classifications, and relations of data products.This will be described in more detail in the next section.

EARTH AND SPACE SCIENCE INFORMATICS AND THE SEMANTIC WEB
ESSI connects the research work of all the different fields of geo and space sciences with outcomes of modern information technology (IT) developments and domain-related geoinformatics research activities.In particular, such ESSI objectives as the generation of metadata extensions of data, the classification of geoscience data and information, the development of concept models for the description and discovery of relations among data product classes, and the usage of taxonomies and ontologies play an important role in on-demand discovery, sustainable management, and systematical use of multi-domain data and information for the generation of knowledge for a better understanding of the complex system Earth.Both the NASA Global Change Master Directory (GCMD) (http://gcmd.nasa.gov),which defines metadata standards for Earth observation data and collects standardized metadata of various Earth science data sets and provides services relevant to global change and Earth science research, and the International Virtual Observatory Alliance (IVOA), which coordinates the development and deployment of tools, systems, and organizational structures for the international utilization of astronomical archives as an integrated and interoperating virtual observatory in the development of standards for the classification of metadata, lead the way in Earth and space science informatics research and development.

Sources of Semantics, Metadata Standards, and Concepts
Semantics, in combination with the Semantic Web, describes the meaning of the content of digital objects in the World Wide Web (WWW).There are many different kinds of digital objects in different file formats on the WWW, such as Web pages (HTML, XHTML), textual documents (ASCII, PDF, MS Word, XML), pictures (GIF, JPG, PNG, SVG), films (MPEG, AVI), sound files (WAV, MP3), and others.There is an object classification related not only to the type but also to the physical format of the object although the information about the objects' semantics in this classification is rather small.There is no detailed information about the content of the digital objects and the relations among these objects available.Structure and sometimes content related information of digital objects often are placed in headers of the digital object files.The following section describes the development of an interoperable standard for the description of semantic information of images.
The very common JPG file format for digital picture objects does not have a formal header structure, but there are functions available to obtain information about width and height (in pixels), the number of components (1 = grayscale or 3 = RGB), and more very technical details.A new standard for pictures taken from digital cameras, the Exchangeable image file format (Exif), (http://www.cipa.jp/english/hyoujunka/kikaku/pdf/DC-008-2010_E.pdf) has been introduced, which provides information or metadata such as date, time, and location of the record, various camera settings, a thumbnail, description, and copyright data.Exif data is embedded within the image file itself.But even now, the semantic information about the content of the picture is very limited.After the introduction of an International Press Telecommunications Council (IPTC) Interchange Model (IIM) (http://www.iptc.org/site/News_Exchange_Formats/IIM/), a set of semantic related metadata attributes is now available not only for images but also for text and other media types.There exists a good deal of content related standardized metadata attributes (IPTC tags), such as headlines (title of a picture), captions or abstracts (description of the content of a picture), keywords (list of subject data by discrete words or phrases that describe the content of a picture), and credits (name of the photographer) provided.This evolutional development of a standard for a semantic description of digital images has been continued by the usage of XML for the definition and introduction of an IPTC core schema and extensions (http://www.iptc.org/cms/site/index.html?channel=CH0099) for the eXtensible Metadata Platform XMP (http://www.adobe.com/products/xmp).
There is a much more complicated picture for digital geoscience data objects.Many geoscience data objects lacking any metadata are on the Web.There are data files with some metadata located at the header of the files, and there are data products, existing as pair or tuple of data file(s) and related standardized and also non-standardized metadata files.Standards for metadata (file) formats are, e.g., Dublin Core for documents (literature), Directory Interchange Format (DIF) for Earth observation data, and ISO 19115 for geographic information.All these standards have common metadata attributes for the "definition" of digital objects related to • identification: (unique) identifier, • content: title, description, keywords, temporal, and spatial coverage, • persons and rights: creator, publisher, citation, • technical data: format, type, language, • cross linking: source, relation, audience, • life cycle: creation date, review date.
Metadata files are often used to describe geoscience data products and product types.In addition, metadata files can be used to describe further metadata categories or classes, such as geoscience projects or missions, platforms and instruments, institutions, and organizations as well as science domains and applications.On the other hand, there are already complete concepts and class models for the description of domain specific geoscience data available, such as Geoscience Markup Language GeoSciML (http://www.geosciml.org)for geoscience information associated with geologic maps and observations and the QuakeML (https://quake.ethz.ch/quakeml)for the representation of seismological data.There are different concepts for the integration of science literature as another source of semantics, such as books, journals, articles, and reprints, based on geoscience data.Often metadata standards provide specific attributes for reference citation rules or the usage of Digital Object Identifiers (DOI) (http://www.doi.org/), which point to a publication's web address using a specific handle service.It is much more difficult to integrate such digital Web objects, such as unstructured and "free" text documents, without references or PowerPoint presentations and posters.
Considering the connection between data and publication from the other side, the important link from geosciences literature to data used in a paper normally does not exist.However, there are ideas to use unique and persistent identifiers, which are valid worldwide, such as DOI or URN (http://www.doi.org/factsheets/DOIIdentifierSpecs.html), for the missing links.Most structured geoscience metadata content is provided by scientists who are experts and are generating both data products and appropriate metadata.On the other end of the semantic spectra is the scientific knowledge in literature.However, in principle, there is much more information and knowledge available in the scientific community that is not used in an efficient way.A solution for this challenge is the usage of modern Web 2.0 feedback techniques, which open a method for qualified collaboration among scientists and other users who work with the data provided by information systems and data centers.Existing Web 2.0 methods provide application for, e.g., tagging activities, social navigation analyses, and collaboration at workspaces, such as Wikis and others.

The ISDC Metadata Class Model
Geoscience data products are the most important digital data objects the ISDC manages and provides to the science community.These data products are classified in product types related to geoscientific domains and sub-domains.Therefore, the ISDC metadata class model is driven by the product type metadata class.Another driver of the classification of the model is the class project (or mission).A scientific project is always a requirement and the frame for the generation and processing of data products.Because most raw data are measured by instruments (or sensors), the class instrument is included in the model.Often instruments are physically mounted on platforms, such as satellites, ships, or observatories.Therefore, it also makes sense to introduce the class platform.In addition to these functional classes, the class institution (or organization) is modeled as its own class in order to encapsulate the object properties of this class.This makes sense because objects of this class often have strong relations to objects of the other classes: project, platform, instrument, and of course, product type.A simplified and schematic class model is provided in Figure 6.The XML schema of the ISDC product type class, which is compliant with the GCMD DIF version 9 schema, is available at the ISDC (http://isdc.gfz-potsdam.de/xsd/base-dif.xsd).A special feature of the ISDC metadata concept is the ability not only to deal with high level classes such as seen in Figure 6 but also to have a concept for the description of geoscience data products.The metadata of data products are necessary for the management of the millions of data products controlled by the ISDC backend and for the insertion of information into the ISDC catalog system for the retrieval component and the product browser of the ISDC.Even though the XML schemata for the data products are derived from the appropriate product type schema, there is a unique schema necessary for each product type.The reason for this diversity lies in the product type dependent characteristic of attributes for a detailed and complete description of geoscience data products.Such specific data parameter attributes are, e.g., the GPS Satellite No. for geoscience data products derived from GPS satellite measurements and, e.g., the reference frame for Earth rotation parameter products.In the data product metadata files, there is also data file-specific technical information, such as file size, MD5 sum, and others.In order to ensure the interrelation between product type and data products of the same product type, some rules and a mechanism for the ISDC are defined.For example, the basic part of the unique identifier (Entry_ID metadata attribute according to the GCMD DIF standard) and the file names for product types and data products are identical.8 illustrate the principles of the ISDC metadata concept for the integrity of product type and data product using the example of product type CHAMP Rapid Science Orbit "CH-OG-3-RSO" and an individual data product "CH-OG-3-RSO+CTS-CHA_2000_219_10" of the same type.Not only are the basic part of the name of the CHAMP Rapid Science Orbit product type and the appropriate data product identical, in this case CH-OG-RSO, but also the basic part of the metadata attribute "Entry_ID" is the same.In addition, there is a specific metadata attribute Parent_DIF (Figure 8) in the data product schema, which points to the appropriate product type, in this example to the CH-OG-3-RSO product type.Most of the semantics-related attributes of the geoscience product types are filled with keywords (words and word phrases) from vocabularies controlled by GCMD for sciences, projects, instruments, platforms, and locations (http://gcmd.gsfc.nasa.gov/Resources/valids/archives/keyword_list.html).In particular, the controlled vocabulary collection of science keywords related to different science domains of the GCMD is an invaluable treasure for the discovery and modeling of semantic relations among different kinds of geoscience product types and data products.
The GGP ISDC as a part of the GFZ ISDC portal is a very good example of the usage of the complete ISDC metadata concept.To understand the following description of the model and features of the GGP object classes, it is important to know that the modeling of the ISDC product type class matches the GCMD standard, whereas the modeling of the ISDC object classes project, platform and instrument, and institution is still in an evaluation phase and only similar to the appropriate classes at the GCMD.In the GGP ISDC object class model, the GGP project and the appropriate GGP project ISDC metadata document are the framework for all project-related semantics, including such attributes as, e.g., • long and short name of the project (unique names), • geoscience objectives of the GGP project (science keywords related to different science domains according to GCMD controlled vocabulary and free-format description), • references to participating institutions and organizations, • references to observatories and stations (platform).The GGP project is based on an international network of GGP observatories or stations (Figure 3).Each of the participating observatories is modeled in an appropriate GGP ISDC platform metadata document.There is content related information about, e.g., • long and short name of the observatories, • observatory characteristics (platform keywords and free-format description), Data Science Journal, Volume 11, 18 October 2012 • location of the observatory (geographical coordinates, data from a gazetteer), • references to scientific instruments, such as superconducting gravimeter and meteorological sensors, • references to institutions and organizations that run an GGP observatory.• MD5 sum: e16d17b525fef9a3b8420f3936ab8ba6 are added to the appropriate metadata schema.Finally, a set of compulsory GCMD DIF standard attributes, such as: • unique identifier: GGP-SG-MIN+ba000100, • title: GGP Minute Data, • keywords: Solid Earth > Geodetics/Gravity > Gravity (controlled vocabulary from GCMD) is included in order to create a complete, and independent from the product type metadata document, viable metadata document.For the first time, the complete ISDC metadata concept is used for the modeling of all semantic aspects of the GGP project.The next section describes the status of the GGP project metadata classes and related documents for an ontology-based approach, according to Semantic Web requirements.

ISDC Ontologies and Web 2.0
The conceptual model of the ISDC metadata classes related to the GGP project is presented in a paper by Ritschel et al. (2008) In order to increase the amount of relevant metadata and to include the knowledge of the geoscientific community, a new prototype of the ISDC portal provides a Web 2.0 interface for the annotation of product types.
Registered ISDC portal users can add keywords and keyword phrases, which are structured in specific keyword classes, such as: • science domain, • project, • instrument (platform + instrument), • product type, • institution, • person, • application.
In addition to the input and creation of new keywords, the ISDC user can define new keyword classes.Both new keywords and keyword classes are checked by a review process.After a positive validation, the new keywords and keyword types become part of the searchable ISDC vocabulary.In a second step, the new keywords, keyword phrases, and classes can achieve a folksonomy level, as described in (http://vanderwal.net/folksonomy.html).New keywords and keyword types also can become part of controlled vocabularies, e.g., GCMD's Science Keywords and Associated Directory Keywords.
Other geoscience semantic information, such as social navigation data, collected and analyzed during the use of the ISDC portal, and information that is provided by geoscience related Blogs, Wikis, and further Web-based communication platforms, can also be used not only for a better integration of metadata within a certain class but also between different metadata classes.

CONCLUSION
Today, there is a need for sustainable handling of geosciences data, information, and knowledge as pointed out, for example, in the Tsukuba declaration (http://egy.org/files/TsukubaDeclaration.pdf) and in Klump et al. (2008).The growing complexity of geoscientific questions, such as the nonlinear feedback mechanism of climate variability, can be faced only with the help of millions of multi-domain data sets.The discovery of appropriate data sets and semantic relations not only within domain-specific classes of data sets but also between multi-domain classes is necessary.Classified and standardized metadata for geoscience data products, product types, instruments, platforms, and projects are the key for the integration of semantic information within metadata classes and among them according to the Semantic Web approach.The GFZ ISDC portal is based on standardized, classified metadata and a semantics-driven conceptual model that has the potential for a sustainable management of geoscience data products and important related information.Using different sources of semantics, such as standardized geoscience metadata from data providers, very different kinds of literature, and the knowledge of the user community, the ISDC concept connects and integrates geoscience data, information, and knowledge.

Figure 2 .
Figure 2. Map of superconducting gravimeter observatories (ggp_stations) of the Global Geodynamic Project

Figure 3 .
Figure 3. Graphical User Interface of the GFZ Information System and Data Center (ISDC) portal At the end of 2008, the ISDC portal was providing almost 300 different product types (a geoscience classification of a set of appropriate data products [data file + metadata document], such as the CHAMP satellite orbit product type containing all data products of the type CHAMP orbit) related to geodesy, geophysics, oceanography, and meteorology, with almost 20 million data products and a total of 12.5 Terra bytes.A daily file flow of approx.5000 data files coming into the ISDC archive system and going out to the users is managed by the ISDC backend components.All data products at the ISDC are processed and provided by 19 national and

Figure 5 .
Figure 5. e-Research Life Cycle, data curation and data related processes* Lyon (2007) (*added by Ritschel, B.)Daconta et al. (2003) describe the problems in terms of informatics: information overload instead of re-usage of data, proprietary systems instead of component based, extensible, and interoperable systems, and poor content aggregation instead of content-interoperable data sources.The e-research life cycle (Figure5), according toLyopn (2007), not only illustrates the life cycle of data from an idea, through the generation of data, to the transformation from data into knowledge but also addresses the data related processes, such as project elaboration and application and modeling.Artifacts of these data related processes are digital sources of semantics for cross-domain descriptions, classifications, and relations of data products.This will be described in more detail in the next section.

Figure 6 .
Figure 6.Simplified ISDC metadata class model, including the ingestion of controlled and free vocabulary Figures 7 and 8illustrate the principles of the ISDC metadata concept for the integrity of product type and data product using the example of product type CHAMP Rapid Science Orbit "CH-OG-3-RSO" and an individual data product "CH-OG-3-RSO+CTS-CHA_2000_219_10" of the same type.Not only are the basic part of the name of the CHAMP Rapid Science Orbit product type and the appropriate data product identical, in this case CH-OG-RSO, but also the basic part of the metadata attribute "Entry_ID" is the same.In addition, there is a specific metadata attribute Parent_DIF (Figure8) in the data product schema, which points to the appropriate product type, in this example to the CH-OG-3-RSO product type.Most of the semantics-related attributes of the geoscience product types are filled with keywords (words and word phrases) from vocabularies controlled by GCMD for sciences, projects, instruments, platforms, and locations (http://gcmd.gsfc.nasa.gov/Resources/valids/archives/keyword_list.html).In particular, the controlled vocabulary collection of science keywords related to different science domains of the GCMD is an invaluable treasure for the discovery and modeling of semantic relations among different kinds of geoscience product types and data products.

Figure 7 .Figure 8 .
Figure 7. Cut-out of the XML schema of the ISDC product type: CHAMP Rapid Science Orbit "CH-OG-3-RSO" . The ISDC concept consists of concept nodes, which are metadata classes or attributes of classes, and the relations between two concept nodes.These relations are designed by concept arrows and corresponding linking phrases, which qualify the direction and the kind of the relation between ISDC metadata classes, e.g., the class: product type "gets data from" the class: instrument.Results of this conceptual model are RDF schema files, N-Triples, and Web Ontology Language (OWL) documents, which contain the semantic information and relations for each class in an integrated form.There are, e.g., OWL files describing the semantics of each ISDC product type of the GGP project.These OWL files now can be used within a framework of reasoning applications for inference and deduction purposes.It is planned to extend the ISDC ontology model by using NASA's Semantic Web for Earth and Environment Terminology (SWEET) ontologies (http://sweet.jpl.nasa.gov/ontology/).Depending on the funding of GFZ ISDC related projects, it is planned to continue the research activities related to ISDC ontologies and to start a technical development for applications of ISDC OWL files in the framework of the ISDC portal.