Proposed design for service oriented architecture applied to gridded data sets

With the escalation of forecast model data (including ensembles), spanning the bridge from research to operations has been a cumbersome path. The length of time it takes to introduce new observations or forecast information to the forecast office community can easily exceed one year. Software infrastructure upgrades and deployment of an upgrade to support new information is costly and the management of data at the Weather Forecast Office has become time-intensive.

In order to reduce the time and effort necessary to transfer scientific research information into operations, the National Weather Service (NWS) has been investigating the use of new technologies for the forecast effort. Efforts to design new infrastructures to support ingest, visualization, information generation, and collaboration functionality of the forecast office are underway. In particular, the concept of services for gridded data sets is under study.

This paper will describe the current investigations of different Web services available in the research community that are exchanging scientific data. The proposal of a functional, service-oriented architecture will be presented. A collaborative development effort between the National Centers for Environmental Prediction (NCEP), NWS, and the Forecast Systems Laboratory (FSL) will also be detailed.