THE JAMSTEC METADATA PUBLICATION AND SEARCH SYSTEM

The Japan Agency for Marine-Earth Science and Technology (JAMSTEC) provides users of its data with comprehensive search services that enable users to find data in JAMSTEC’s various data dissemination sites. These are the “JAMSTEC Data Search Portal” that helps users to search for observational data on a map and the “JAMSTEC Data Catalog” that enables users to find data sites by selecting science keywords. The “Data Search Portal” and the “Data Catalog” have been developed and operated as dedicated metadata publication and search services that collaborate with data sites in JAMSTEC.


INTRODUCTION
JAMSTEC has performed a wide variety of observations and research on both the ocean and the Earth.In 2007 JAMSTEC established a data policy (http://www.jamstec.go.jp/e/database/data_policy.html) on the handling of data and samples from its research activities and developed rules and management systems based on this policy.JAMSTEC operates over 100 research cruises a year and disseminates observational data from these cruises in the "Data Research System for Whole Cruise Information in JAMSTEC" (http://www.godac.jamstec.go.jp/darwin/e/), images in the "E-library of Deep Sea Images" (http://www.godac.jamstec.go.jp/jedi/e/index.html),rock sample information in the "GANSEKI" (http://www.godac.jamstec.go.jp/ganseki/), sediment core sample information in the "Core Data Site" (http://www.godac.jamstec.go.jp/coredata/e/), and biological sample information in the "Marine Biological Sample Database" (http://www.godac.jamstec.go.jp/bio-sample/index_e.html).Additionally, various research projects operate their own data dissemination sites.
As the data management system in JAMSTEC was developed, the numbers of data sites and the amount of opened data increased rapidly.Also, users had to search for data in various data dissemination sites designed for the specific data type, and there were many inquiries to the data management office concerning whether the specific data was opened or not and how to find the data.Therefore, JAMSTEC has developed two data search services that enable users to find data by themselves easily.One such service is the "Data Search Portal" (http://www.godac.jamstec.go.jp/dataportal/index_eng.html), which helps users to search for observational data by specifying the area of interest on a map.The other is the "Data Catalog" (http://www.godac.jamstec.go.jp/catalog/data_catalog/index_en.html), which leads users to the appropriate data dissemination sites by selecting science keywords.Figure 1 shows the schematic flow chart in finding data using these two services.Both systems provide just search service based on the metadata.Users are able to learn the details of the data on the linked data dissemination page and, if needed, to download the data or apply for the use of off-line data.

2
DATA SEARCH PORTAL

Objective
The search target of the "Data Search Portal" is an individual observational datum from a research cruise, mooring, or terrestrial observation.This portal was developed to serve as a comprehensive data search service through various data dissemination sites by specifying data types and the area of interest.The "Data Search Portal" has been in operation since November 2008.

Method
With the "Data Search Portal", users are able to confirm the distribution of observations of specific data types and then select the area for search.If needed, additional criteria (research vessel name, cruise period, observation variables, etc.) can be added.The list of search results includes the common metadata (date, location, data id, etc.) and a link to the data dissemination site that enables users to move directly to the relevant data page.This is a simple URL link, and the "Data Search Portal" can cooperate with any data system as far as a data page can be specified with a URL.
The expected major users of the "Data Search Portal" are scientists who are searching for data in the same area at different times or other data types in the same area.Therefore, the authors developed a map-based spatial retrieval function using a web GIS server (ArcIMS 9.2).

Metadata
Location information in points or lines is extracted from data files and merged into shape files with observation metadata.These metadata include several common metadata for cruises or dives (cruise number, period, observation variable, etc.) and optional metadata for a specific data type (chief scientist, water depth, area name, etc.).The vocabulary of the observation variables is determined by the frequently observed variables in JAMSTEC.Data types in the "Data Search Portal" come mainly from research cruises or dives, but some data types come from mooring observations, ocean bottom stations, and terrestrial observations.By the end of 2011, over 30,000 observations, shown in Table 1, were opened on the "Data Search Portal". 3 DATA CATALOG

Objective
While the target of the "Data Search Portal" is an individual observation, the target of the "Data Catalog" is a larger data dissemination unit (data sites, databases, datasets, etc.).JAMSTEC aims to provide a comprehensive metadata publication and search service for all of the data in JAMSTEC using these two systems in complement.
The "Data Catalog" is also designed to be a data publication system for scientists who do not have their own data dissemination sites.JAMSTEC opened the "Data Catalog" in September 2011.

Method
The user interface of the "Data Catalog" is a classification tree of hierarchical keywords.Expected users of the "Data Catalog" are scientists searching for data not in their main research field.These scientists are probably not familiar with the appropriate keywords to use when searching for data in other fields.The classification tree confirming the structure of the keywords makes finding an appropriate keyword easy for scientists from other disciplines.
The hierarchical keywords of research fields or observations are listed in the category tree.Users are able to narrow down the list of metadata by selecting keywords in the category tree.Going down the tree, related sub categories and metadata of the keyword will appear as shown in Figure 2.
The content of each data site is not fully described in the common metadata format.Metadata in the "Data Catalog" is used only for finding data sites.Details of the data site and data download depend on the each data site.Users are also able to search for metadata by using free text.

Figure 1 .
Figure 1.Schematic flow of data search using "Data Search Portal" and "Data Catalog"

Table 1 .
Data types and number of observations on the "Data Search Portal"