The Environmental Data Server (EDS) provides homogenous access to meteorological and hydrodynamic data via web services. The EDS provides visualization services (WMS) as well as data access services so model data can be used in decision support applications such as Search & Rescue and Oil Spill response. It is designed to be scalable and modular and utilizes a Service Oriented Architecture (SOA). Users (clients) access the available data by using a web service, and the web service returns data based on the request. The web service may return data to the client from data products stored in distributed servers, or a more common approach is that the catalog server retrieves data to a central server. A TDS instance is layered on top of the EDS Data Store to provide catalog and DAP services. The EDS consists of three major sub-systems governing data acquisition (the Catalog Server), data storage (the Data Store), and data request (the EDS web services).
Several GLOS data products are currently served by EDS including the NOS Currents, NAM Winds, and GLERL Currents, Ice, Water Level, and Waves products. Data provided by GLOS partners is harvested by RPS’s production EDS instance. The production EDS is currently used in the following manner:
- A source for gridded (models, radar, satellite) wind and current data from global, national and regional providers
- Model data are stored in its native format (generally NetCDF) and grid structure. See example of unstructured FVCOM for Gulf of Maine being displayed with EDS services. Model data are not regridded on the EDS servers
- Data are stored and maintained in its native format and can be delivered with custom data services
- Aggregation services are available to provide temporal and, in some cases, spatial aggregation products
- The EDS manages structured and unstructured gridded data and data are made available via visualization and data access web services
- Visualization services can be integrated into map-based web sites managed by different entities. New visualization techniques are being developed including high performance “on the fly” contouring techniques – see example below. Visualization is performed directly from the model grids, no data conversions are needed
- Staging and development versions of EDS are used to evaluate and distribute pre-operational model products. This provides a framework for easily sharing experimental model results.
- Model providers do not need to make any changes to their workflow – the EDS simply collects model results as soon as they are available in their native format.
- RPS is working on a new set of unstructured grid access services so unstructured gridded data may be subsetted and regridded
- RPS also maintains deep archives of the data. RPS archives most source data from the point that the new data source was added to the Coastmap EDS. The archived data are available for restoration to Coastmap Production servers with very little lead-time.
- Redundancy – RPS hosts 3 instances of the EDS, Development, Staging, and Production with the ability to quickly add instances. RPS also deploys an instance of the EDS on the Amazon Cloud.
- The EDS currently collects a wide variety of federal and regional data that can be made available to users and the infrastructure
- The EDS currently focuses on delivery of 2D time-varying data, but development is ongoing to efficiently deliver subsetted 3D data
- A new model skill assessment project is underway with the US Coast Guard to automate model evaluation based on surface drifter comparison
