導入氣象觀測資料於GPS網型定位高程精度之研究

Abstract

GPS信號由衛星至地表經過大氣層導致信號的折射及延遲，在對流層當中GPS電磁波之傳遞速率，由於L1及L2兩種頻率的信號之折射量相同，所以GPS觀測時，無法利用雙頻觀測消除折射量，也因此，估計對流層延遲量是否精確，成為了影響GPS高程定位精度的最大因素之一。 處理對流層延遲量的方法，包含氣象數學經驗模式、對流層參數估計、外部觀測資料改正。在靜態網型定位，通常會以氣象數學經驗模式配合對流層參數估計法來處理對流層延遲量。內政部於雲林縣北港站設有一台水氣微波輻射儀，可觀測此區域對流層延遲絕對量，配合對流層參數估計法可精準預估北港站附近短基線各站的延遲相對值。如此可結合外部改正配合參數估計，探討在台灣濁水溪沖積扇地層下陷區域，對靜態網型定位之坐標趨勢影響。 本研究處理之GPS數據，其觀測期間為2007年11月2日至2008年12月31日，共426天。成果顯示，加入水氣輻射儀資料後，在地層持續下陷情況較明顯的KTES站及TKJS站，高程方向標準差由2.088公分及2.675公分降低至1.092公分及1.321公分。

GPS signals are transmitted from satellite to earth through the atmosphere, causing refraction and delay. The tropospheric layer, which belongs to neutral atmosphere, the transmission speed of electromagnetic wave changes barely with signal frequency. For this reason, the troposphere bias cannot be eliminated with dual- frequency of carrier phase observation. Also, the troposphere bias becomes one of the most significant factors which influence the accuracy of GPS height positioning. To reduce the troposphere bias, the strategy, meteorological experience models, parameter estimation, and external correcting is frequently used. For the static network computation, empirical meteorological models are usually combined with parameter estimation. On the other hand, the direct meteorological observations may be used for improving the positioning accuracy. This scheme is referred as correction with external observation. The Ministry of Interior has installed one WVR (Water Vapor Radiometer) at PeiKang GPS Station, which can observe the absolute troposphere bias, and predict precisely the relative troposphere bias of each short baseline station nearby PeiKang when using parameter estimation collaboratively. This study investigates the potential of correction with external meteorological observations. From observing the time series of coordinates in a static network, experiments are conducted. This GPS network was established in Choshui River alluvial fan of Taiwan for subsidence monitoring. The data set includes 426 days GPS daily observation from 2007/11/2 to 2008/12/31. The results show that the standard deviation of vertical direction are from 2.088cm to 1.092cm at KTES station and from 2.675cm to 1.321cm at TKJS station with WVR observation.