弱綜觀環境下台灣午後對流特徵及其客觀預報

Abstract

摘 要 台灣地區在弱綜觀環境條件下的暖季，午後對流(afternoon thunderstorm, TSA)為其主要的天氣型態，而其準確的時空預報極具挑戰性。由於雷達資料為本研究主要定義與特徵化午後對流之資料，為避免非氣象回波影響本研究分析結果，尤其在具複雜地形的台灣地區，故其品質控制(quality control, QC)為首要進行目標。本研究利用三年(2005–2007)的氣象局屬四座雷達回波資料建立其氣候統計資訊並結合地形以及其各掃描策略，所架構之各雷達具氣象條件影響之合成仰角(hybrid scan)，經評估可有效去除受地形雜波(ground clutter)及地形遮蔽(beam blockage)效應所影響之回波資料。 利用上述經QC的雷達資料，以及雲對地閃電、衛星、探空和地面氣象站等資料，探討在弱綜觀環境條件下共四年(2005–2008)的暖季(5–10月)午後對流特徵。在此期間，平均時雨量顯示其最大降雨量發生於1500–1600 LST (local standard time)間，而降雨的最大發生頻率則是侷限於平行於山脈走向的低斜坡帶上。進一步利用經QC的雷達資料將弱綜觀環境再區分為有發生午後對流(TSA days)及無發生午後對流(non-TSA days)兩類，其地面平均風向、溫度及露點溫度均呈現明顯差異。在TSA days期間，其對流前環境(preconvective environment)於近地面層具有相對較non-TSA days而言較暖(約0.3–2.4 °C)及較濕(約1.6–2.8 °C)的環境。探空除了顯示整層也有較為濕暖及較小風速的環境條件外，於850–650 hPa層間，露點差異在TSA days與non-TSA days間，可高達3.0–4.5 °C。 這些對流前的環境差異(預報因子)，可做為模糊邏輯算則(fuzzy logic algorithm)之隸屬函數(membership function)，提供TSA的客觀預報。由地面測站與探空資料所得的預報因子之特徵函數，可強調在TSA days與non-TSA days期間的對流前特徵差異。本研究建立的客觀預報TSA之fuzzy logic方法，整合28個隸屬函數，並根據最佳化臨界成功指數(critical success index, CSI)得其權重。這些權重可定量描述對流發生前相對較暖濕的綜觀環境、海風傳送水氣至內陸，以及內陸的弱風速等，皆為提供有利發展午後對流活動的條件。除此之外，持續性預報法(persistence forecast)亦可視為預報TSA的參考預報因子。經兩年(2009–2010)暖季(5–10月)之個案進行驗證並評估，顯示此客觀預報TSA的fuzzy logic方法可整合弱綜觀環境所呈現之各項對流前環境參數特徵，提供TSA的機率性預報，未來可進ㄧ步運用於作業中。 關鍵詞：回波氣候統計，合成仰角，午後對流特徵，對流前環境特徵，模糊邏輯算則，隸屬函數。

Over Taiwan island with complex terrain, during the warm season (May–October) under weak synoptic-scale forcing, afternoon thunderstorms are the main weather system, and it is still faced the rigorous challenges to accurately forecast the thunderstorms on small temporal and spatial scales. In this study, quality control of radar reflectivity data is the primary procedure to accomplish since radar is the mainly dataset for identifying and characterizing afternoon thunderstorms (TSA). Three years’ of radar reflectivity data from four radars from 2005 to 2007 are analyzed and the reflectivity climatology is developed. The climatology is applied in the construction of climatology–based hybrid scans to minimize the impact of ground clutter and beam blockages. The spatial and temporal characteristics and distributions of thunderstorms in Taiwan during the warm season (May–October) from 2005–2008 and under weak synoptic-scale forcing are documented using radar reflectivity, cloud–to–ground lightning, satellite, radiosonde, and surface observations. Average hourly rainfall amounts peaked in mid-afternoon (1500–1600 local standard time, LST). The maximum frequency of rain was located in a narrow strip, along the lower slopes of the mountains and parallel to the orientation of the mountains. Significant diurnal variations are found in surface wind, temperature, and dewpoint temperature between days with and without afternoon thunderstorms (TSA and non-TSA days). Before thunderstorms occurred, on TSA days, the surface temperature was warmer (about 0.3–2.4 °C) and the surface dewpoint temperature was moister (about 1.6–2.8 °C) than non-TSA days. Sounding observations from northern Taiwan also showed warmer, higher moisture, and smaller wind–speed conditions on TSA days relative to non-TSA days. The largest average difference was in the 850–650 hPa layer with values of 3.0–4.5 °C drier on non-TSA days. And, wind speeds on non-TSA days between near surface and 6 km were 2–3 m s–1 stronger than on TSA days. A fuzzy logic algorithm has been proposed to provide the objective guidance for the prediction of TSA using these membership functions of preconvective factors. Features functions of thunderstorms are derived from surface stations and sounding measurements that highlight the preconvective characteristics of days with TSA that best differentiate them from synoptically undisturbed days. There are total 28 membership functions and associated weights in fuzzy logic approach are derived and determined based on the optimization of the critical success index (CSI). The results quantitatively illustrate that synoptically relatively warmer and moister conditions, sea breezes transport moisture into the inland, and weak winds in the inland provide the proper conditions in inducing afternoon convective activity. In addition, the relatively simple persistence rule is also useful for prediction of afternoon thunderstorms, i.e., there is high probability that afternoon thunderstorms occur consecutively for a few days when the environmental conditions are right. Evaluation with the validation dataset (2009–2010) suggests that the fuzzy logic algorithm has the capability to integrate the preconvective factors and provide the probability guidance for the predictions of TSA, and the highly predictive potential that could be implemented in real-time operations.