Abstract: To accurately and intelligently find data and information among massive polar science datasets, this paper proposes a linked data method for polar science data searching. Firstly, a systematic study of the characteristics of polar science data was conducted to discover the implied semantic information of these datasets. Secondly, a polar science linked data index system containing three aspects and four-level indices was built. Thirdly, a linked calculation method of the essential characteristics of polar science data was discussed in detail. In general, essential data characteristics include temporal, spatial and content elements. Finally, we designed and implemented a prototype system for polar science linked data discovery based on the Polar Science Data Center of the National Earth System Science Data Sharing Platform, which is one of the 23 platforms of the National Science and Technology Infrastructure. In an experimental environment, it uses the metadata of polar science datasets to abstract the corresponding index value, and then it calculates the degree of correlation among different datasets. Thus, a polar science linked data network can be built automatically. The results showed that by using the polar science linked data method, the prototype system can retrieve not only needed data, but also potential linked data. Such linked data method research can be further used to integrate, fuse, and share polar science data.

Key words: Polar, Linked, Semantic network, Data sharing