open access

https://doi.org/10.14943/doctoral.k16256

https://hdl.handle.net/2115/97674

雷を直接計算する気象モデルを用いたダウンバーストに先行する雷活動の激化に関する数値的研究

英語

2025-03-25

Hokkaido University

157p

Mixed-phase clouds largely characterize severe weathers such as heavy rain, lightning, downbursts, and so on. They contain cloud microphysical processes involving both liquid and solid particles. In this doctoral dissertation, the mechanism for the increase in lightning activity preceding downbursts associated with mixed-phase clouds was investigated by using a meteorological model coupled with a bulk lightning model (BLM) that explicitly calculates lightning processes. The validity of cloud microphysics schemes in the model was also evaluated through comparison with in-situ ground-based measurement and satellite observation. In Chapters 2 and 3, the evaluation of bulk cloud microphysics schemes used in the model was conducted through comparison with observational data. Three bulk microphysics schemes—two 1-moment bulk schemes that predict mass and diagnose number concentration of hydrometeors, and a 2-moment bulk scheme that predicts both mass and number concentration of hydrometeors—were evaluated through comparison with the observation data of a disdrometer measurement and Himawari-8 satellite. The evaluations were conducted for a snowfall event in Hokkaido. We here applied a newly developed method based on the machine learning algorithm to the data. The results of the evaluations clarified that the 2-moment bulk scheme better reproduced the particle size–fall velocity distribution observed by the disdrometer, in addition to the radiance and brightness temperatures retrieved from a satellite. In Chapter 3, the evaluation for a 2-moment bulk scheme was extended to snowfall events during a winter season in Hokkaido. To analyze the measurement data for the whole winter, a method to automatically classify precipitation particle types was developed with an aid of machine learning algorithms. The method was developed by applying self-organizing maps to the particle size–fall velocity relationship data, which were estimated using the expectation-maximization algorithm based on particle size–fall velocity distribution data measured by disdrometers. This novel approach enables the automatic classification of precipitation particles without any prior data, and additionally, the 2-moment bulk scheme was evaluated by utilizing this approach. In Chapter 4, the rapid increase in lightning activity preceding the downbursts in multi-cell convective clouds was investigated using the 2-moment bulk scheme, which was validated in Chapters 2 and 3. The model successfully simulated the rapid increase in lightning activity preceding a downburst. Our analysis elucidated that the transition in the convective cells from tilting to upright under specific conditions of vertical wind shear and cold pool strength is critical for the occurrence of the rapid increase in the lightning activity. The insights of this doctoral dissertation would contribute to deeply understanding of the dominant cloud microphysical processes in mixed-phase clouds that characterize the severe weathers. These insights are expected to improve the warning for downbursts and to sophisticate meteorological models. In the future, further understanding of mixed-phase clouds is anticipated through meteorological models and extensive observational data.

2025-03-25

甲第16256号

博士(理学)

(主査) 准教授 佐藤 陽祐, 教授 稲津 將, 教授 見延 庄士郎, 准教授 佐々木 克徳, 室長 橋本 明弘 （気象庁気象研究所）

理学院（自然史科学専攻）

博士論文

Version of Record