臺灣大學: 地理環境資源學研究所朱子豪古卡茲Kuckartz, JoostJoostKuckartz2013-03-212018-06-282013-03-212018-06-282011http://ntur.lib.ntu.edu.tw//handle/246246/247990Tropical cyclones are one of the most devastating weather events on earth. Much effort has gone into the prediction of its path and intensity but sudden changes are not yet possible to predict. A definition for a sudden track change is presented by statistical evaluation of historical tropical cyclones of the West-Pacific basin, done by three-point extrapolation and real versus extrapolation differences. Nine historical tropical cyclones are selected for evaluation of MTSAT-1R geostationary satellite imagery toward sudden track change indicators. Animations are created and evaluated for the detection of external forces on tropical cyclones and temporal data is extracted in a new extraction technique to find possible precursor indicators of sudden track change. Indicators of a sudden track change are validated by using the method on eight tropical cyclones between 2007 and 2009. Two programs, one focused on GIS for the statistical evaluation and one focused on remote sensing for image transformation and data extraction, are developed to assist in the process. For four tropical cyclones main airflows are evaluated and the temporal data extraction technique is demonstrated on the tropical cyclone’s cloud distribution for finding indicators of a sudden track change. The results of this research can assist in a better understanding of the influence of clouds on tropical cyclones and the programs are developed for easy adjustment in future research.127985556 bytesapplication/pdfen-US颱風Tropical CycloneHurricaneTyphoonSudden ChangeBest-TrackTurn DetectionSpatio-Temporal Data ExtractionSpatio-Temporal Data EvaluationSatellite Imagery EvaluationMTSAT-1Rthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/247990/1/ntu-100-R97228025-1.pdf