以TaiwanVVM大渦模擬系集建立嘉義地區午後對流之預警前兆

Abstract

嘉義地區夏季午後對流常迅速發展，對飛航安全構成潛在威脅，常伴隨能見度驟降、跑道積水與風切等現象。觀測與研究指出，該區對流系統受地形加熱、海陸風輻合與中層風場調節等因素共同影響。然而，現行預報作業對對流觸發與演變機制掌握不足，難以因應具操作性的短時預警需求。本研究運用有高解析度(500 m)臺灣真實地表狀況與地形高度資料之渦度向量方程雲解析模式(TaiwanVVM)，分析62組弱綜觀條件模擬中嘉義地區的午後對流發展過程，並建立一套分類與診斷架構。首先以嘉義機場方框內降雨強度達3 mm/hr且覆蓋範圍逾13 km²為條件，篩選出44組具降水事件的模擬個案，進一步針對其中15組850 hPa為西南風背景者進行詳細分類。根據模擬對流觸發與移動特性，本研究將其歸類為「單觸發」與「雙觸發」兩類型事件。前者由海風與坡面風輻合所觸發，對流活動集中於午後初期、壽命短、降雨熱區偏嘉義平原西側；後者則於午前至午後初期生成第一波對流，午後再因冷池推展誘發第二波對流，降雨範圍向東延伸至丘陵地帶，呈現2–3小時的雙階段演變結構。診斷分析指出，700 hPa中層風場為分類關鍵：穩定西南風有利海風推進並對應單觸發事件；而雙觸發事件常伴隨中層弱風或東風，有利冷池發展與午後再對流。此外，雙觸發事件中的冷池強度與推展範圍普遍高於單觸發事件，並與中層回流層結構共同幫助對流維持與提供再觸發的條件，顯示此類午後對流發展需整合冷池、風場與海風推進等多項因素進行判識。本研究據此建構具物理意義之對流分類架構與早期診斷指標，整合海風推進深度、中層風場型態與冷池特性，可應用於嘉義機場午後對流之短時預警作業。研究結果補強作業經驗模型之物理依據，亦提供後續發展即時守視工具的基礎。未來建議納入東風背景或無降水個案，並結合雷達、衛星與地面觀測進行驗證，以提升模式應用廣度與實務預測效能。

Afternoon convective systems in the Chiayi region often develop rapidly during summer, posing potential threats to aviation safety, such as sudden drops in visibility, runway flooding, and wind shear. Observational and modeling studies have shown that the development of these systems is jointly influenced by terrain-induced heating, land–sea breeze convergence, and mid-level wind modulation. However, current operational forecasts have limited capability in capturing the triggering mechanisms and evolution patterns of convection, making it difficult to support short-term early warning needs. This study analyzed the simulations of the TaiwanVVM model—a cloud-resolving model based on the vector vorticity equation and configured with 500-meter resolution and realistic Taiwanese topography and land conditions. From the 62 simulated cases under weak synoptic condition, the objective is to reconstruct the evolution of afternoon convection over the Chiayi region and establish a physically grounded classification and diagnostic framework. Using a threshold of rainfall intensity ≥ 3 mm/hr and coverage ≥ 13 km² within the Chiayi Airport domain, 44 convective cases were identified. Among these, 15 cases with 850 hPa southwesterly background flow were selected for further classification. Based on convective initiation and propagation characteristics in the simulations, cases were categorized into two types: single-trigger and dual-trigger. The former are initiated by convergence between sea breezes and upslope winds, with short-lived convection occurring during the early afternoon and rainfall concentrated over the western Chiayi Plain. In contrast, dual-trigger events exhibit a first convective episode before or shortly after noon, followed by a second wave of convection in the afternoon triggered by cold pool propagation, with rainfall extending eastward toward the hill regions and a typical interval of 2–3 hours between the two stages. Diagnostic results reveal that the 700 hPa wind field serves as a key classification factor. Stable southwesterly flow favors inland sea-breeze penetration and is associated with single-trigger events, while dual-trigger cases commonly occur under weak or easterly mid-level flow, which promotes cold pool development and secondary convection. Moreover, dual-trigger events generally feature stronger and more extensive cold pool structures, which, with the return flow layer at mid-levels, enhance convective maintenance and re-initiation. These findings suggest that the dual-trigger events need to be interpreted through the combined influence of cold pools, wind fields, and sea-breeze dynamics. Based on these results, this study proposes a physically meaningful classification system and early-stage diagnostic indicators that integrate sea-breeze depth, mid-level wind patterns, and cold pool characteristics. The framework has potential applications in short-term convective nowcasting for the Chiayi Airport. It also strengthens the physical foundation of empirical operational models and supports the future development of real-time monitoring tools. Future work may extend to include easterly background flow or non-precipitating cases, and incorporate radar, satellite, and surface observations for further model validation and enhancement of practical forecasting capabilities.