Hydrologic Ontology Workshop

A Hydrologic Ontology workshop was held at the San Diego Supercomputer Center, headed by Michael Piasecki (Drexel Univ) on January 27-28, 2005.

Summary for Hydrologic Ontology Workshop:

San Diego Supercomputer Center
January 27-28, 2005

List of Participants:

  1. Michael Piasecki, Drexel University (convener)
  2. David Maidment, University of Texas, Austin
  3. Thanos Papanicolaou, University of Iowa
  4. Edwin Welles, NOAA, National Weather Service, OHD
  5. Luis Bermudez, Monterrey Bay Aquarium Research Institute (MBARI)
  6. llya Zaslavsky, SDSC
  7. Kai Lin, SDSC
  8. Ashraf Memon, SDSC

The above group met on January 27 and 28 at the SDSC to discuss the development of a hydrologic ontology that can be incorporated into the GEON framework. The intension of creating this ontology is based on an ongoing research effort to develop a prototype Hydrologic Information System (HIS) as part of the Consortium for the Advancement of the Hydrologic Sciences, Inc. (CUAHSI) initiative funded by NSF. The HIS prototype team is working closely together with the SDSC to develop a digital library system (DSL). In fact, one of the components of the funded effort is to establish a formal interface between the HydroViewer system (DSL) and the GEON system. The HIS effort, however, does not have an identified task that would permit the expenditure of a substantial effort in terms of funds and time to develop an ontology framework that could be registered with the GEON system. It is clear that GEON is based on an innovative concept that has the potential to providing the GeoSciencs with an umbrella framework. Hence the hydrologic community should undertake an effort to register a specifically designed ontology framework with GEON that would enhance its interoperability with neighboring specialty areas like geology and atmospheric sciences beyond what is currently being envisioned in the CUAHSI-HIS project.

The workshop was centered on the objective of discussing and subsequently defining how an upper hydrologic ontology could look like. This upper ontology would serve as a backbone for other more detailed ontologies that would link into the backbone. Several alternatives were discussed during the course of the workshop. One of them was to divide the hydrologic realm vertically into an atmospheric layer, a surface water layer and a sub-surface (groundwater) layer. Another alternative discussed was to identify spatial features along which the water travels, for example a vertical water budget. Other discussions centered on the idea to use a taxonomy of hydrologic terms to construct the hydrologic ontology. Another alternative that was evaluated was to sub-divide the hydrologic realm into a small number of key areas of interest and then expand these key areas with an appropriate number of detailed ontologies to incorporate various aspects and concepts.

During the discussions it became clear that there is no clear "winner" as all approaches have a certain number of pros and cons. However two alternatives emerged as likely candidates for a backbone upper hydrologic ontology. The first alternative is to use a "top-down" structure that centers on the concept of a measurement, regardless where it occurs. The basic idea is to view the hydrologic realm as collection of data measurements that are necessary to describe hydrologic phenomena. The second idea is to sub-divide the hydrologic realm into areas of typical interest. These are: models (analysis, prediction), data (observations, measurements), processes (phenomena), human interaction (flooding, WQ, ...), and representations (maps, time series ...). These two concepts are distinctively different in nature and it would remain to be researched which of them could better serve as the Upper Hydrologic Ontology.

As a result of the deliberations the group identified the following key research tasks:

  1. Development of a higher level Hydrologic ontology based on the above mentioned concepts. The group felt no clear affinity for one or the other concepts. As a result, two top ontologies may need to be developed and placed next to each other. Depending on the task at hand a user may use either one of them to address the objective.

  2. Development of lower ontologies that can be merged with the top ontology.
    a.   development of ontologies from database schema (like ARCHydro and the NWIS data base) via XML schema libraries
    b.   development of a processes (or phenomena) ontology
    c.   development of modeling ontology

  3. Development of a well defined Hydrologic Controlled Vocabulary that can be used to query the hydrologic realm. One suggestion made was to use common queries as a starting point to identify important aspects in the taxonomy of the CV.

There were a number of smaller tasks that surfaced during the discussion like the development of a generic time series ontology. In addition, the group spent some time to identify key questions that could be used to guide the development of the ontology. It became clear however that the area of Hydrology is so complex that there are an infinite number of questions that could be asked. As a result, this approach (though recommended when developing an ontology) was viewed as inefficient or not suitable for identifying the scope of an upper or any sub-level ontologies in the provided workshop setting. However, the issue of formulating a set of competency questions that the ontology framework needs to satisfy should not be disregarded and may need to be visited in the course of the ontology development. Finally, the workshop group used the Protégé 2000 ontology tool to develop a first version of a time series description based on the time series approach used in ArcHydro. This exercise proved valuable as it showed the possibility to move existing schema into ontology format giving rise to the believe that this can in fact be done on a much broader scale.