On the Forecast of Summer Afternoon Short-Duration Thunderstorm over Taipei Basin.

碩士 === 國立臺灣大學 === 大氣科學研究所 === 107 === This study aims for a better objective forecast for the Summer afternoon short-duration thunderstorm over Taipei basin. Based on the thermodynamics of convection, this study utilizes factors from surface observations (Lin et al., 2012; 周等, 2016) and information...

Full description

Bibliographic Details
Main Authors: Nai-Chen Liao, 廖乃臻
Other Authors: 郭鴻基
Format: Others
Language:zh-TW
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/h2dn8a
Description
Summary:碩士 === 國立臺灣大學 === 大氣科學研究所 === 107 === This study aims for a better objective forecast for the Summer afternoon short-duration thunderstorm over Taipei basin. Based on the thermodynamics of convection, this study utilizes factors from surface observations (Lin et al., 2012; 周等, 2016) and information from upper-air radiosondes. This study focuses on the afternoon thunderstorm cases occurring in July and August from 2012 to 2017. Among the total 372 days, 223 days are considered as undisturbed cases including 33 days of extreme precipitating cases (Ex-PR, the station at Taipei basin and valley which has hourly rainfall over 40 mm); 81 days of precipitating cases (PR, the station at Taipei basin and valley which has hourly rainfall over 0 mm but less than 40 mm), and 109 of non-precipitating cases (No-PR, the stations at Taipei basin and valley are all without rainfall). Those days affected by fronts, tropical cyclones (TCs), tropical depressions (TDs), and external and/or large-scale weather systems are not considered. Surface moisture flux convergence (MFC) and other observational factors for weather forecasting are calculated and analyzed in these 223 days. Comparing the forecasting factors and the occurrences of afternoon thunderstorms, K-Index, Total Precipitable Water and moisture in mid-level (700 hPa) show better Threat Scores (TSc.) than surface observations. However, insignificant differences were found among MFC in Ex-PR, PR, and No-PR cases.