An ocean of data
NOAA begins integration work for ambitious observation program
- By William Jackson
- May 15, 2008
THE NATIONAL OCEANIC and Atmospheric Administration is launching a pilot program to make data from thousands of ocean observations more accessible to scientists and analysts who use the information to predict severe weather and other maritime conditions.
'We're establishing interoperable access to existing databases,' said NOAA Corps Capt. Chris Moore, deputy director of the Integrated Ocean Observing Program. 'It is predominantly linking legacy systems together.'
The initial phase of the program will be Internet-based, using open standards to make data about five core variables ' seawater temperature, salinity, sea level, currents and ocean color ' gathered by NOAA and other organizations available in a common format that multiple programs can easily use. As the program advances, additional variables will be added, and program managers plan to design a robust communications architecture to serve users.
'Our efforts will reduce the steps' needed for extracting and using data, Moore said. 'We haven't unlocked the full potential of this information.'
The Integrated Ocean Observing System (IOOS) is a cooperative program of 17 federal agencies and 11 regional partners that collect data from ocean observation systems in coastal and nearby waters. Although a coordinated effort began in the 1990s, IOOS was launched in February 2007 with NOAA as the lead agency. The first goal in NOAA's strategic plan is to improve access to data from multiple sources through a data integration framework.
The pilot phase will standardize data in online databases from NOAA's National Data Buoy Center, Center for Operational Oceanographic Products and Services, and Coastal Watch Program.
Four NOAA decision support tools will use the data to predict hurricane intensity, coastal flooding and harmful algal blooms and make integrated ecosystem assessments.
Although these NOAA systems also gather data from other sources, the initial phase focuses on databases the agency regularly contributes to and uses. The initial set of variables and modeling areas was selected because 'addressing everything at once [involved] too many moving parts,' Moore said.
The data integration program will use Web services and standards developed by the Open Geospatial Consortium, an industry standards organization of more than 350 companies and universities that generate, use or create software for the data.
'We define standards to enable the easy movement of data from one source to another,' said OGC spokesman Sam Bacharach.
Geospatial data ties information about conditions to a specific location. There is a rapidly growing market for geospatial data for scientific uses, such as IOOS and business decision-making, and consumer applications, such as online mapping and direction services.
'We're defining an open way to get and use data regardless of the source,' Bacharach said. The organization has published 25 standards ranging from the simple, such as Web Map Services for making requests in an image, to the complex, such as the Geographic Markup Language, which can define any geospatial phenomenon in a way an application can understand.
'We chose an OGC set of Web services because it represented a wide set of open standards,' Moore said. 'We are looking for something that is flexible and extensible across a wide variety of systems.'
The standards define not only the format of the data but also surrounding metadata and transfer protocols, which help the system integrate data that different organizations collect for their own purposes. NOAA's National Data Buoy Center, for example, gathers data from 90 buoys and 60 automated fixed stations along with another 70 to 90 observations made by satellites, high-frequency radar stations measuring surface currents, National Estuarine Research Reserve Stations, Army Corps of Engineers wave sensors and Acoustic Doppler Current Profilers installed on offshore oil rigs by the federal Minerals Management Service.
'There is a lot of work to get data from one source to work with data from another source,' Moore said.
IOOS is developing the application scheme and implementing OGC Web services for the initial data sources, and it will expand the system to include other nonfederal data sources during the next year. In 2009, program managers will identify additional decision support tools that will use the integrated data, and developers will begin system design engineering for a more comprehensive integration program. The development phase of the data integration framework is expected to be completed in 2010.
William Jackson is freelance writer and the author of the CyberEye blog.