Towards Linked Data Conventions for Delivery of Environmental Data Using netCDF

Abstract

Part 2: Information Systems, Information Modelling and SemanticsInternational audiencenetCDF is a well-known and widely used format to exchange array-oriented scientific data such as grids and time-series. We describe a new convention for encoding netCDF based on Linked Data principles called netCDF-LD. netCDF-LD allows metadata elements, given as string values in current netCDF files, to be given as Linked Data objects. netCDF-LD allows precise semantics to be used for elements and expands the type options beyond lists of controlled terms. Using Uniform Resource Identifiers (URIs) for elements allows them to refer to other Linked Data resources for their type and descriptions. This enables improved data discovery through a generic mechanism for element type identification and adds element type expandability to new Linked Data resources as they become available. By following patterns already established for extending existing formats, netCDF-LD applications can take advantage of existing software for processing Linked Data and supporting more effective data discovery and integration across systems

netCDF-LD SKOS: Demonstrating Linked Data Vocabulary Use Within netCDF-Compliant Files

“…While Yu et al [5] proposed and Baird et al [6] have demonstrated sophisticated integration of Linked Data into netCDF files, we have tested simples of integrations only whereby we place Universal Resource Indicator (URI) links [8] to SKOS concepts within attribute metadata in netCDF files. Our reason for taking this very simple approach is to enable the demonstration, and ultimately the adoption, of some limited Linked Data and Semantic Web functionality with the lowest possible barrier to uptake.…”

“…The netCDF specification includes a set of software tools used to manipulate netCDF files [6]. These tools allow people to view netCDF metadata and data, create netCDF files and view visual representations of netCDF's gridded arrays.…”

“…Some examples of tools, their developers and their usages are given in Table 1. Unidata, the makers of netCDF Command-line utility converts netCDF data to human-readable text form (CDL or XML) ncgen 6 Unidata A program that creates a netCDF dataset fro CDL input ncview 7 Scripps Institution of Oceanography A netCDF visual browser The netCDF Operators toolkit 8 A community of developers Command-line programs that take netCDF, HDF, and/or DAP files and derive new data, compute statistics, print or otherwise manipulate them In order to make our simple SKOS-only deployment of netCDF-LD usable by people without detailed knowledge of Linked Data, we have created a prototype Python 9 utility called ncskosdump which wraps the ncdump tool (see Table 1) and adds a series of options for the retrieval and display of Linked Data that the tool is able to extract from a netCDF file, including links to vocabulary terms as described in Section 2.…”

“…Sophisticated methods for the inclusion of Linked Data within netCDF files have been demonstrated at a previous ISESS conference called 'netCDF-LD' [6] and an example of netCDF-LD using Linked Data reasoning has also been given recently [7]. This paper is both a 'next step' to those bodies of work and also a step back with respect to the complexity of implementation.…”

“…Demo vocabulary online at http://pid.geoscience.gov.au/def/voc/netCDF-LD-st-demo5 http://www.unidata.ucar.edu/software/netcdf/netcdf-4/newdocs/ncdump-man-1.html6 http://www.unidata.ucar.edu/software/netcdf/netcdf-4/newdocs/ncgen-man-1.html 7 http://meteora.ucsd.edu/~pierce/ncview_home_page.html 8 http://nco.sourceforge.net/9 The Python programming language, v2.7: https://www.python.org/download/releases/2.7/…”