Junshi Ito is Associate Professor in the Atmospheric Science Group, Department of Geophysics, Graduate School of Science, Tohoku University, Japan, since 2020. He received his Ph.D. from the University of Tokyo in 2010. His research focuses on high-resolution numerical weather simulations, including large-eddy simulations, to investigate meso-scale and micro-scale atmospheric processes. These simulations are used to study fundamental atmospheric dynamics and severe weather phenomena, such as heavy precipitation associated with linear-shaped precipitating systems, tropical cyclones, and tornadoes, in both idealized and realistic settings. His research group also develops long-term regional reanalysis datasets over Japan and conducts statistical analyses of various meso-scale weather systems.
Numerical Simulation and Statistical Analysis of Senjo-Kousuitai
Quasi-stationary, quasi-linear heavy precipitation systems, known in Japan as Senjo-Kousuitai, are examined using fine resolution numerical simulations and a long-term regional reanalysis dataset. We focus particularly on back-building convective systems, which are generally difficult to predict because convection self-organizes and sustains the system.
To clarify their characteristics, we performed downscaling simulations for several observed cases of Senjo-Kousuitai associated with back-building convection (Oizumi et al., 2018, 2020; Ono and Ito, 2024). In addition, idealized experiments were conducted to isolate fundamental mechanisms by eliminating the complex factors in real cases (Ito et al., 2021). These experiments successfully reproduced stationary linear systems and enabled investigation of detailed structures, formation mechanisms, resolution dependence, and sensitivity to vertical wind shear. At large eddy simulation resolution (dx = 100 m), cumulonimbus clouds are shown to be periodically generated from thermals in the boundary layer, contributing to the maintenance of the stationary system.
We also use a long-term regional reanalysis dataset over Japan (RRJ-Conv; Fukui et al., 2024), which provides annual occurrences of these events since 1958. Although its horizontal resolution (dx = 5 km) is marginal for explicitly resolving Senjo-Kousuitai, the dataset captures synoptic environments favorable for their formation. While individual reproduced systems do not necessarily match observed events, the reanalysis indicates that Senjo-Kousuitai-like systems have occurred in earlier decades when observational records were limited.