Semantic web holds promises for ocean observing needs

Ocean observing has been gaining the interest of federal and local policy makers in recent years, partly because scientific data is now clearly telling a story that garners public attention. The health of our oceans is critical to the stability of many ecosystems, and is impacted on the physical, chemical as well as biological fronts. Ocean health has a profound impact on coastal habitats, marine life and human well being. Working to create the ability to monitor the conditions of our coastal waters, the ocean observing community has been implementing a national integrated ocean observing system (IOOS). While significant progress has been made, there are many technical issues remaining to be addressed to ensure that critical data can be found accurately and timely. Ranking among the challenging technical issues such as inter-region interoperability, archival and sharing of data, is data discovery. Unfortunately, data discovery has been taking a back seat to other more prominent issues and has not received the attention it warrants. Looking toward the future, Semantic Web is an emerging technology that holds promise for addressing important data discovery issues facing the ocean observing community. The current approach for solving the data discovery of a regional IOOS typically involves a text based search engine to index the observation sensor metadata and subsequently serve up results. Unfortunately, traditional search engine capabilities are insufficient for ocean observing data because they were built originally for finding unstructured textual information on the World Wide Web. While they do well for keyword search strings such as "ocean\´08 conference information", they are quite inadequate for searching against structured data such as "time- series ocean surface temperature or salinity levels in the Gulf of Mexico". Traditional search engines deploy various complex algorithms, take into account the characteristics of the HTML tags on a web page, analyze the l- inks and their significance based on its source, and the occurrence frequency of the sought keyword(s) in the web page itself. Ocean observation sensor data are non-textual information and typically stored in flat files or kept in relational databases, and are not suited for presenting in HTML web pages. This has been a challenge for those that work with non-textual scientific data and information expressed in numbers and patterns of numbers such as drug discovery, healthcare, and ocean harmful algal bloom density. In the ocean observation environment, before user can access the data, sensor output information typically goes through a process of aggregation, quality control and mediation. Many of the data management and communication functions that involve ingesting inbound data are machine-to-machine based communication. Semantic Web is an environment created specifically with syntax for machine-to-machine communication and which naturally understands data context. In the name of inter-regional interoperability, the regional ocean sensor data environment is shaping up with many similarities to semantic Web including the employment of uniform resource identifier (URI), extensible markup language (XML), resource description framework (RDF), and other elements of a syntax based computing environment. With the emerging semantic Web technologies, instead of serving up pointers and layers of pointers to ocean sensor data as it does today, users can phrase their request in human natural language and let the semantic Web and its inference engine intelligently analyze and supply the appropriate response. The semantic Web is not a new Web but it is rather a natural extension of the World Wide Web per Berners- Lee, the founder of the Internet; it has the potential of revolutionizing the way we look at the oceans and, in effect, allows us to protect our endangered environment and dependent fragile ecosystems.