Issue link: https://read.dmtmag.com/i/50311
'Rigor Spectrum' Expanding at Both Ends OPENGEOSPATIAL CONNECTION T BY DAVID ARCTUR he rapid rise of Web mapping, lightweight apps, and visual representations of spatial data generated by search engines sug- gests that geospatial rigor (data quality) and exper- tise don't matter so much any more. In fact, although more categories of people who aren't "GIS experts" can indeed get remarkably useful results using mass- market, spatially enabled Web services (and that trend will undoubtedly continue), technology advances are revealing require- ments for greater technical rigor and expertise as well as better standards. Earth Isn't Round? Consider that global climate models (GCMs), unlike GISs, often don't account for different models of Earth's shape (datum). A geodetic datum is a set of reference points and an associated model of Earth that provides an origin and orientation for mapping coordinate systems. Most GCMs are developed on the assumption that Earth is perfectly spherical, which simplifies their already complex equations. However, Earth is closer in shape to an ellipsoid (a flattened sphere), and many GIS datasets are ref- erenced to an ellipsoidal datum (such as WGS84, the datum for GPS receivers worldwide). In preparing input data from GIS sources for GCMs (e.g., land cover and elevation), scientists often don't perform coordinate transformations to convert the input coordinate values from ellipsoidal to spherical datums. Likewise, most GIS users of GCM outputs often fail to convert from sphere to ellipse. Until now, information about the coordinate David Arctur is director of Interoperability Programs, Open Geospatial Consortium; e-mail: darctur@ opengeospatial.org. 30 reference systems (datum or projection) hasn't been retained in either step. Positional errors resulting from confusion over the datum are on the order of 20 kilometers at mid-latitudes. This wasn't an issue when GCMs ran at coarse spatial resolutions, but improvements in processing speed and the resolution of input data enabled improving spatial resolution in climate models. As GCM resolutions now approach 10-50 kilometers, the positional shift due to datum becomes significant. GEO W ORLD / D ECEMBE R 2O11 A figure shows the spatial resolution of climate models used in periodic assessment reports by the Intergovernmental Panel on Climate Change from 1990 to 2007. The need for a solution increases with the growing diversity of data used as inputs to climate models and the growing number of GIS users who want to use climate-model outputs. Listen to Experts How do the climate-science community and others ensure reliable and accurate communication of location from differing climate models, devices, services and platforms? Each community needs to convene their experts to discover answers. The solution may involve best practices using existing standards or require stan- dards that don't yet exist (e.g., standards for carrying provenance data through complex workflows). Existing, widely implemented OGC standards don't by themselves solve all problems. Much work is being done by various geoscience communities and other interest groups in OGC to develop application schemas, workflows, modeling-framework interfaces and best practices based on existing standards. Progress is driven by committed representatives of communities that focus on unmet requirements for sharing or integrating geospatial information—rep- resentatives willing to engage in a collaborative and inclusive search for shared solutions. Author's Note: Please send any comments to info@opengeospatial.org. FIGURE REPRINTED WITH PERMISSION FROM RICHARD C. J. SOMERVILLE AND SUSAN JOY HASSOL, PHYSICS TODAY, VOL. 64, PAGE 48, 2011. COPYRIGHT 2011, AMERICAN INSTITUTE OF PHYSICS.