Press Releases
JAMSTEC
KYOTO UNIVERSITY
Development of a New System (SEAL) Enabling the Efficient Search and Acquisition of Large-Volume Climate Simulation Data
1. Key points
- ◆
- A new technological system (SEAL) was developed that enables the rapid and efficient extraction of necessary information from several petabytes of large-volume climate simulation data.
- ◆
- The system uses only approximately 1% of the time required to extract the necessary information relative to conventional systems. It also employs an interface that can be operated intuitively and without specialized knowledge.
- ◆
- It is expected that SEAL will greatly contribute to future research, such as climate change projections and impact evaluations, and advance examinations of region global climate change adaptation strategies.
2. Overview
Yujin Nakagawa from the Japan Agency for Marine-Earth Science and Technology (JAMSTEC) and Koji Koyamada of Kyoto University have developed a new technological system that enables the rapid and efficient extraction of information from several petabytes of large-ensemble climate data (*1) under the Social Implementation Program on Climate Change Adaptation Technology (SI-CAT) commissioned by the Ministry of Education, Culture, Sports, Science and Technology (MEXT). An analysis of the Database for Policy Decision Making for Future Climate Change (d4PDF) is crucial for performing climate change projections and evaluating the associated uncertainties quantitatively. However, the volume of the d4PDF is enormous, as it comprises three petabytes of data.
To date, conventional systems have performed searches using information acquired before performing simulations, such as model and variable names. This creates substantial obstacles, requiring several hours for searching and acquiring information, as well as necessitating a high-capacity recording device. Thus, the physical quantities obtained after a simulation, such as precipitation and temperature, were combined by prefecture and then temporally compressed to develop a new technology that enables the efficient searching and acquisition of necessary data. The system equipped with this technology, the System for Efficient Content-Based Retrieval to Analyze Large Volume Climate Data (SEAL; Fig. 1), is made available on the Data Integration and Analysis System (DIAS) developed under a MEXT-commissioned project.
Compared to conventional web-based search systems, using SEAL reduces the size of the data downloaded by a user to less than 0.5% and the total time required to find necessary data to less than 1% (Fig. 2).
Therefore, it is expected that this system will help advance research concerning factor analyses on past notable meteorological events, uncertainties in future climate change projections, and impact evaluations (Fig. 3). Additionally, SEAL is expected to be widely used not only by researchers but also by ministries and agencies, local governments, and industries, owing to its user-friendly interface.
This research was published in the academic journal Progress in Earth and Planetary Science issued by the Japan Geoscience Union on February 27, 2020 (JST).
- Research Institute for Value-Added-Information Generation, JAMSTEC
- College of Humanities and Sciences, Nihon University
- Academic Center for Computing and Media Studies, Kyoto University
- Research Institute for Global Change, JAMSTEC
- Laboratory of Meteorology, Hokkaido University
- Meteorological Research Institute, Japan Meteorological Agency
- *1
- Ensemble climate data:
Data created with multiple calculations by adding slight variations to the initial value upon performing a future climate simulation.
Figure1 Conceptual diagram of SEAL.
Figure2-1 Screenshot of the web-based user interface of SEAL. The users can select common conditions of retrieval (names of datasets, experiment types, physical variables, and retrieval types) in the area surrounded by dashed line (a). In addition, the users can also select and input unique conditions of retrieval associated with the retrieval types in the area surrounded by dashed line (b). Areas surrounded by dashed lines (c) and (d) show explanation of the retrieval types and supplements of the input fields, respectively.
Figure2-2 An example of a histogram generated by SEAL-V. The histogram shows a distribution of averaged daily precipitations (mm/day) using the +4K future climate simulations for Tokyo metropolis.
Figure3 Schematic of the expected advantages of using SEAL. The elapsed times shown in the figure are actual measured values for the +4K future climate simulations based on download speed from DIAS to JAMSTEC (4–120 Mbps).
Contacts
- (For this study)
- Yujin Nakagawa, Project Senior Engineer, Information Engineering Program, Research Institute for Value-Added-Information Generation, JAMSTEC
- (For press release)
- Public Relations Section, Marine Science and Technology Strategy Department, JAMSTEC