應用TRMM衛星資料進行陸上颱風降水反演及其降雨特徵之研究

• Main Authors: Tsai Ming-Da, 蔡明達
• Other Authors: 陳萬金
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/89008378291057779169

博士 === 國防大學理工學院 === 國防科學研究所 === 98 === Taiwan is hugely threatened by typhoons during the summer and fall seasons annually. On one hand, copious rainfall brought by typhoons often leads to enormous economic losses and damage. On the other hand, rainfall brought by typhoons is the primary source of fresh water in Taiwan. Using satellite microwave observation to rainfall retrievals on Taiwan land could improve the accuracy of the rainfall estimation in the mountainous areas, where rain gauge observation is not available and radar observation might not be accurate enough. It will help provide more useful information to make better decisions on typhoon warnings and disaster prevention. To manage water resource in advance and to prevent human lives and properties from damaged by flooding and mudslides, both the forecast of the typhoon movement and the typhoon rainfall estimation are very important and necessary. In this study, data from the Tropical Rainfall Measuring Mission (TRMM) satellite's Microwave Imager (TMI) and the rainfall observation by the Automatic Rainfall and Meteorological Telemetry System (ARMTS) were employed to quantitatively estimate rainfall using the scattering method for land rainfall retrievals over Taiwan. The radiation from the land is stronger and more complicated than that of the sea, so that a pixel mixed with ocean and land areas could not have accurate identification of rainfall. Therefore, the TMI data were used to analyze Taiwan’s land surface area to identify the land parts in which the contamination of ocean, having low radiation, was avoided. In addition, data from Automatic Rainfall and Meteorological Telemetry System and TRMM/PR (Precipitation Radar) were employed to analyze the characteristics of land rainfalls over Taiwan. At last, the satellite rainfall retrieval algorithms for TMI observation have been developed for different rain types, classified from PR, to improve the accuracy of the rainfall estimation. The Scattering Index over Land (SIL) equation in Taiwan area was established in this study. According to the statistical analyses of the SIL values with no-rain status, the rainfall threshold of 8 K was determined. Finally, the rain retrieval rate algorithms were set up by using coincident data pairs of SIL and rain gauge. Based on the rain types determined by TRMM/PR, the rainfall areas in Taiwan land were divided into convective rainfall areas and stratiform rainfall areas with/without bright band. Analyses in these different rainfall areas were conducted to make sure whether this classification could be used to improve the accuracy of quantitative rainfall retrievals. With the observational data from ARMTS set as the baseline, the verification of the accuracy of the satellite rainfall estimates was performed by comparing the discrepancies between the rainfall estimates that had and had not undergone rainfall type classifications. The absolute mean rainfall error of the convective rainfall areas was reduced from 6.2 to 1.7 mm/hr and from 10.7 to 1.8 mm/hr for bright-band stratiform rainfall areas. To obtain more fine resolution of the satellite rainfall estimation, the Polarization Corrected Temperature (PCT) method was used to conduct the Taiwan land rainfall retrievals. The rainfall threshold of the brightness temperature was found to be 270 K for Taiwan area after the establishment of the PCT equation and the statistical analyses of PCT values of no-rain areas. The relationships between the rain rates and PCT values were then established for different rain types and the satellite rainfall estimations were determined by the above relationships. The results show that the rain rate estimated by PCT was more accurate than that by SIL in convective rainfall areas. The improvement in the accuracy of rain rate estimation was noticeably. For the satellite land rain rate estimation, it is recommended that the combination usage of using PCT method in convective rainfall areas and using SIL method in stratiform rainfall areas is one of the best rainfall estimation strategies.