摘 要 台灣地區夏季期間，常有劇烈午後熱對流的發生，這種對流系統發生的機制，是氣象學家及預報人員相當感到興趣的議題之一。由於午後熱對流發展快速，單單從傳統的觀測資料是不容易做到事前的預警，因此往往造成人、事、物的損傷。尤其台灣地區四面臨海，建立衛星觀測資料之應用能力更顯得重要。本研究主要目的是，建立一套適合台灣附近區域使用的NOAA(National Oceanic and Atmospheric Administration)繞極衛星大氣溫濕垂直剖面反演方法，並將其結果應用在夏季午後熱對流發生之有利大氣條件分析。 在大氣溫濕剖面反演結果顯示，NOAA衛星反演850、700、500、400及300 hPa大氣溫度（露點）的均方根誤差及相關係數分別為2.47（4.43）、2.18（3.13）、1.76（5.07）、1.97（5.73）、2.05 °C（3.73 °C）及0.83、0.82、0.71、0.76、0.67。在溫度方面，雲覆蓋量小於50%的情況下，各層均方根誤差平均約2.0 °C，標準差約0.5°C，在露點方面，雲覆蓋量小於50%的情況下，各層均方根誤差平均約4.3°C，標準差約0.9°C。 本研究利用衛星資料計算台灣地區附近大氣穩定指數及500 hPa水氣含量，並將其結果應用在實際的個案分析之中，結果顯示：大氣水氣含量的多寡及穩定度的狀態，對夏季午後對流系統的生成與發展同樣的重要。此外藉由2003年6 ~ 9月441個樣本分析，計算出大氣穩定度配合水氣資訊所訂定預測對流降水的門檻值，將此門檻值應用於非2003年的470個獨立個案，也有相當不錯的結果，初步的分析結果顯示，其定量預報對流降水準確率可達約85%，此結果似乎可以提升午後熱對流的預報能力。 Intense convections often occur over the Taiwan area during the afternoons of summer. The relevant mechanisms behind this weather phenomenon is one of the various issues that strongly attracts the attention of meteorologists and forecasters. As the afternoon convection development is very rapid, data observed by traditional methods alone can not produce an early warning, which often results in serious damages to the environment and livelihood of people. Due to the lack of traditional radiosonde observational data near the waters of the Taiwan area, it further emphasizes the importance of the applications in satellite monitoring. The major purpose of this research is to establish a method for the polar-orbiting NOAA-15 satellite in retreiving temperature and dew point profiles over the Taiwan area, then we apply this result in convention environment during the afternoon in summer. The results of retrieval using temperature and dew point profiles show that the root mean square errors and correlation coefficients of temperature (the dew point) are 2.47 (4.13) 、2.18 (3.13) 、1.76 (5.07) 、1.97 (5.73) 、2.05 (3.73) and 0.83、0.82、0.71、0.76、0.67. About temperature, the r.m.s. error in each layer is nearly 2.0 degrees when the cloud coverage is less than 50% and the standard deviation is about 0.5 degree. Besides, about dew point, the r.m.s. error in each layer is nearly 4.3 degrees when the cloud coverage is less than 50% and the standard deviation is about 0.9 degree. In this study, we utilize the NOAA satellite data to estimate the atmospheric stability around Taiwan area and moisture at 500 hPa and also apply the results to real cases. It shows that the same importance of both atmospheric moisture content and stability in convention system during the afternoon in summer. Additionally, by using the samples form 2003 June to September (441 cases), we predict the threshold of convective precipitation from atmospheric stability and moisture data and apply to the 470 cases of 2002 also shows good agreements. In elementary analysis, we can get accuracy about 85 % of quantification prediction in precipitation. It shows that the results can improve the prediction of convention system during the afternoon in summer.
