Japan Geoscience Union Meeting 2022

Presentation information

[E] Oral

A (Atmospheric and Hydrospheric Sciences ) » A-AS Atmospheric Sciences, Meteorology & Atmospheric Environment

[A-AS05] Weather, Climate, and Environmental Science Studies using High-Performance Computing

Mon. May 23, 2022 9:00 AM - 10:30 AM 106 (International Conference Hall, Makuhari Messe)

convener:Hisashi Yashiro(National Institute for Environmental Studies), convener:Takuya Kawabata(Meteorological Research Institute), Tomoki Miyakawa(Atmosphere and Ocean Research Institute, The University of Tokyo), convener:Koji Terasaki(RIKEN Center for Computational Science), Chairperson:Tomoki Miyakawa(Atmosphere and Ocean Research Institute, The University of Tokyo)

9:45 AM - 10:00 AM

[AAS05-04] An Ensemble Prediction Experiment for the Extreme Rainfall Event in July 2018 with 500-m Resolution Throughout Western Japan

★Invited Papers

*Kenta Sueki1,2, Yoshiyuki Kajikawa1,3, Sachiho Adachi1,2, Hirofumi Tomita1,2, Satoru Oishi1,3 (1.RIKEN Center for Computational Science , 2.RIKEN Cluster for Pioneering Research, 3.Research Center for Urban Safety and Security, Kobe University)

Keywords:Heavy rainfall event, High-resolution ensemble simulation, Fugaku

In the beginning of July 2018, many areas in western Japan were subjected to extremely heavy rainfall. More than 200 people were killed during this event, which has been recorded as one of the worst casualties caused by a single rainfall event in the past 30 years in Japan. In the present study, we investigate the predictability of the extreme rainfall during this event by conducting a high-resolution ensemble prediction experiment using the supercomputer Fugaku. We used a non-hydrostatic atmospheric model SCALE-RM. Using a one-way online nesting system, we performed the simulations with three different horizontal resolutions of 18 km, 3 km, and 500 m. The innermost domain with a resolution of 500 m covers almost all area of western Japan. The target of prediction is the 72-h total precipitation from 1200 UTC 4 July to 1200 UTC 7 July in 2018. We performed the simulations from three different initial times of 0000 UTC 2, 0000 UTC 3, and 0000 UTC 4 July to investigate the lead time of the 72-h precipitation forecast. The initial ensemble analysis fields of the atmosphere for each initial time were generated by using the SCALE-LETKF system. The ensemble size was 50. The boundary conditions for the outermost domain were generated from the data of NCEP GFS corresponding to each initial time. First, we checked the average 72-h precipitation in western Japan. According to the JMA data, the analysis precipitation was 225 mm. While the ensemble mean of the 72-h precipitation for the forecasts starting from 2 and 3 July was 130 mm and 113 mm, respectively, that starting from 4 July was 163 mm. For the former 2 cases, there were no members whose total precipitation exceeded 200 mm. On the other hand, some members in the latter case predicted the precipitation exceeding 200 mm. This is because the center of main precipitation was slightly farther away from the western Japan for the predictions starting from 3 July and earlier, resulting in less precipitation in the target region. This shows that the risk of heavy rainfall in western Japan had been increasing over time for a few days before this event occurred. We also focused on the Hiroshima area where many sediment disasters occurred during the event. The analysis precipitation was 415 mm in the Hiroshima area. On the other hand, the ensemble mean of the 72-h precipitation in the same area for the initial time of 2, 3, and 4 July were 108, 93, and 199 mm, respectively. Only the ensemble for the initial time of 4 July included some members whose total precipitation exceeded 400 mm. This implies that the risk of disaster in this area increased drastically on 4 July.