XRAIN WebDB System Providing Flexible and Fast Access to Big Streaming Data of Precipitation

The X-band MP radars network, XRAIN, has been deployed around major urban and potentially hazardous areas in Japan. It observes precipitation every minute with a mesh resolution of 250 m mesh. The data are useful not only for national-level analysis, such as runoff and meteorological analysis, but also for making our lives easier and safer with various applications that use the data of a certain region. The example applications are as follows: rainfall notification and taking in the washing, reporting cancellation of a concert, forecast of floods on roads, and suggestion of routes to avoid rainfall. It is also expected for urban inundation management at municipalities and facility managers that an application alerts high-risk torrential rain by calculating the data from radars. However, even if users do not need precipitation information of the wide area in these situations, they have to receive the streaming data collectively as a large binary including the wide area or as a wide map image showing rainfall. These distribution methods that broadcast the entire observation area data can complicate application processes used in everyday life as well as render application creators' interpreting the data difficult. In order to meet various users' needs who depend on local precipitation information, it is necessary to design a Web API system that can provide local precipitation data as a shape of meshes required by an individual user and create its platform that can store streaming big data rapidly and quarry the data flexibly from its database. Therefore, we implemented an XRAIN WebDB system, which stores precipitation radars' streaming big data rapidly and has Web API enabling its users to select data in various shapes by expressing a query as an enumeration of composite rectangles to facilitate mashups on the Web. We developed a prototype application using the Web API to evaluate its design. Furthermore, we experimented on the speed of store and selection using the XRAIN WebDB system deploying real XRAIN data to evaluate the mechanism of the system. The results demonstrate that the XRAIN WebDB system can positively contribute to diverse applications using precipitation radar data in everyday life.