Data interoperability requires agreements on metadata, data models, and service interfaces. Recent developments in standards for geographic information offer considerable potential for discovery and exchange of earth-related information. The ISO Technical Committee 211 is developing a raft of specifications for digital geographic information, metadata and services. In addition, the OpenGIS Consortium is developing a number of implementation specifications for web-enabled rendering and delivery of earth-related data.

We present here a new data model and XML markup language for the climate sciences, based on emerging ISO standards. A number of climate-science 'feature types' are defined: for point, profile, and gridded data, as well as series of these in time and space. The model is applied, as part of the UK NERC DataGrid project, to a wide range of atmospheric and oceanographic data - both observed and simulated. The model is implemented using the current draft of ISO standard 19136 (Geography Markup Language), and drawing on a number of standard conceptual schemas. A mechanism is included for wrapping and aggregating file-based data storage (e.g. netCDF, GRIB, NASA Ames formats) to provide a uniform semantic interface to climate science data.