Relating observations to long term averages using Linked Data

The Bureau of Meteorology publishes a range of observational data from rainfall to river levels. The primary way of accessing these data is through HTML pages or a range of CSV and JSON documents. There are a number of challenges in accessing these data: formats and access mechanisms are different across each data product; there are no explicit relationships between concepts within the data (e.g. whether the specific phenomenon being measured is the same in two different data sets) and no explicit relationships between data sets (e.g. manual daily rainfall observations relate to other rainfall observation types). Linked data is an emerging data-publishing concept that promotes publishing data with links between concepts described in a consistent way. It has potential to assist with issues of identity of objects on the web (e.g. monitoring stations), relationships between data, and standardising access mechanisms. We are publishing near real-time weather observations as Linked Data to allow connections to be made between current conditions and long-term averages. This builds on previous work that published the Australian Climate Observations Reference Network - Surface Air Temperature (ACORN-SAT) data as Linked Data. The published observations allow navigation and query across the two data sets, allowing dynamic comparison of current observations with long-term averages. We describe the vocabularies used to model these data, an approach for linking between related concepts, and the prototype APIs to access the data.