ACHIEVING INTEROPERABILITY IN GEOSCIENCES

Introduction

Much has been said in the geoinformatics community about the need for interoperability. There are many scientific research centers compiling significant data holdings, and many of these recognize the need to bridge between multiple scientific disciplines. For example, at a recent NSF workshop, Building a National Geoinformatics System (NSF 2007), the following was stated by Chris Paola, director of the National Center for Earth-surface Dynamics (NCED): "Important, and often costly decisions concerning land management, restoration, and subsurface resources rely on outdated science and reasoning by analogy rather than process-based analysis. The necessary science comprises elements of geomorphology, ecology, hydrology, sedimentary geology, engineering, social sciences, and geochemistry, but is not any one of these. The foundation of a useful science of Earth-surface dynamics must be synthesis across disciplines and scales, and quantitative prediction."

Similarly, D.A. Miller of the Center for Environmental Informatics at Penn State describes the "Critical Zone, defined by the outer limits of vegetation and the lower boundary of ground water, [which] reflects a complex interplay between the physical, chemical, and biological realms and has become the focus of research among a community of scientists derived from disciplines including, but not limited to, ecology, soil science, biology, geochemistry, hydrology, and geomorphology. A currently funded NSF investigation at Penn State is building tools and infrastructure to promote interdisciplinary research in a developing consortium known as the Critical Zone Exploration Network (CZEN). CZEN is envisioned as a network of sites, people, tools, and ideas…"

These are just two of a very broad and impressive set of research initiatives including EarthRef, GEON, NEON, and many others that are, within their respective communities, seeking to bridge the differences in classification systems and ontologies, semantics, spatial-temporal scales and reference systems, tools, processes, and other barriers that have inhibited interdisciplinary collaboration and integrative research. This is clearly a substantial effort for any one research center or consortium to undertake, and it is not yet complete. The dream and goal remain to carry out integrative studies across multiple scientific data centers and consortia, comprising the broadest possible range of disciplines. This is essential if we are to understand, for example, the causes and implications of climate change. Facing and coming to terms with the barriers to interoperability will also lead to more useful and robust spatial data infrastructures (SDI) among regional, national, and global agencies, which will, in turn, greatly improve the ability of our governments and other organizations to respond to natural and human-caused disasters and emergencies. It is no exaggeration to say that the stakes are high, and that no collection of geosciences research data can be considered exempt from the need to become interoperable with other scientific data for the purpose of interdisciplinary, integrative analysis.

Steps Toward Interoperability

The particular challenge here is to provide for unfettered exchange of information among a great many scientific endeavors, while still accommodating the unfettered requirements of investigators to organize and present their data as per the standards of their own research communities. Fortunately, such advances are being made. Many consortia have adopted the cause of interoperability within their subject domains. A growing number of research centers are striving now to stimulate interdisciplinary research. Bodies of research data that can be compiled through substantially automated means, such as with satellite imagery, synthetic radars, sensor networks, and so on, are often designed around large-scale database architectures that lend themselves to interoperability. Various common schemas based on XML are being developed by different consortia, each with a specific application focus to meet their varied needs. For data with location and temporal content, a growing number of community-level schemas are based on ISO DIS 19136 Geography Markup Language (GML, see http://www.opengeospatial.org/standards/gml), such as GeoSciML (https://www.seegrid.csiro.au/twiki/bin/view/CGIModel/GeoSciML, see also NFS 2007, pp.71-73) and ClimateML (http://ndg.nerc.ac.uk/csml/).

Common sets of interacting web services are also emerging that enable users to find and use the data they seek. The Open Geospatial Consortium (OGC, http://www.opengeospatial.org), which developed GML, has also developed open and international specifications for online data catalog services, web mapping services (to transfer data as simple graphic images), web feature services (for scalar and vector data), web coverage services (for gridded or field-type data), and services for sensor data. What is important about these specifications are that they define interfaces for data exchange; they do not require restructuring of existing databases nor changes in custodial policy.

Functional interoperability among web services is of limited use unless the data provided by these services can be understood by the greater audience of users and accommodated by an array of client software programs. In the context of the complex, multi-faceted investigations that one encounters in interdisciplinary work, it is essential to integrate semantics across multiple disciplines. Fortunately, there has been much progress in recent years in developing the means for expressing semantic content, and in its application by a great many scientific and other information communities. Well-defined, widely-adopted ontologies now exist or are emerging in many disciplines.

But this is not an easy process. Cultivation of a productive consortium and engagement with the right stakeholders is not assured. The OGC however, has had a number of successful collaborative experiments with its Interoperability Program. Since its first Interoperability Initiative in 1999, the OGC has evolved a process that confronts precisely the challenges outlined in this paper. The process to date has enabled collaborations of diverse stakeholders, mash-ups of data from multiple different sources, software development by teams of skilled programmers (sometimes from competing software vendors working together), and complex analyses using data from fields as diverse as atmospheric science, toxicology, hydrology, geology, and marine science.

The OGC Interoperability Institute (OGCII) became operational in 2006 to promote the use and benefits of the continuing development of interoperable geoprocessing with the broader scientific community as well as public sector agencies and research organizations. In order to accomplish these objectives, OGCII works closely with the OGC Interoperability Program to involve researchers in both public and private sector in testbed and pilot activities which develop the techniques and specifications which enable interoperability.

It should not be necessary to mobilize vast new funding initiatives to address this issue; considerable funds are already in place for developing integrative, interdisciplinary datasets, tools, and processes. What is needed is the recognition by both the research centers and the national agencies of the importance of semantic as well as functional interoperability, and support for ongoing collaboration to achieve this end.

References

NSF 2007, Building a National Geoinformatics System: A Community Workshop, March 14-15, 2007, Denver, URL: http://www.geoinformatics.info/Featured%20Events/BUILDING%20A%20NATIONAL%20GEOINFORMATICS%20SYSTEM.doc