增益相對重力測量精度之研究

Abstract

本論文研究三種主題關於相對重力測量的精度。第一個主題是相對重力網的點位精度，以內政部今年執行之一等重力測量為例。不同的網形結構有不同的變方協變方矩陣和重力點精度。較理想的網形是要考慮網形精度和經濟因素。約制15個絕對重力點會提升重力精度達41％。在山區相對重力測量設計以「虛擬觀測」方式來提升網形精度。第二個主題是提出三種方法來決定相對重力儀的漂移：1.單點重複觀測。2.測線多餘觀測求漂移係數法。3.重複點多餘觀測求漂移係數法。以第2種方法效果最好，處理漂移原理如下：在網形平差中所有測線上重力儀的漂移首先求得，再利用此漂移的影響對測線的重力觀測量做改正。本實驗是用第2種方法以Lacoste Romberg G型和Lacoste Romberg EG兩種相對重力儀所施測過的重力網形來做測試，G型所施測過的重力網形精度提升0.5∼1.3 ，EG型所施測過的重力網形精度提升2.3∼3.3 。第三個主題是海潮負載的影響。用FG5絕對重力儀來觀測重力網中某個站的時間序列重力值，在利用最小自乘調和分析法從時間序列重力值中求出 分潮負載。 分潮負載可用來改善相對重力測量海潮負載的影響。利用這種方式來改善彰化和雲林重力網，精度提升0.1∼0.3 。

Three issues concerning the accuracy of relative gravimetry are studied in this work. The first issue concerns point gravity accuracy in a relative gravity network. Simulated data from the Taiwan first-order gravity network of Ministry of the Interior are used for investigating this issue. Different configurations of network result in different variance-covariance matrices and in turn point gravity accuracies. An optimal network is obtained by considering accuracy and economic factors. Fixing gravity values at 15 absolute gravity stations improves point gravity accuracy by 41%. In mountain areas, the network accuracy can be improved by adding “pseudo observations”. The second issue concerns the drift of a relative gravimeter, which is determined by three methods: (1) repeat observations at a fixed point, (2) redundant observations along a survey line, and (3) redundant observations at repeat points. The second method yields the best result. An approach dealing with drift is that, in a network adjustment, the drifts of the gravimeter along all survey lines are first estimated by the second method, and the observations along survey lines are corrected for the drift effects. The results from two experiments using this approach and LCR G and LCR EG types of gravimeter at gravity networks show that, the accuracy of estimated gravity value is improved by 0.5~1.3 (G type) and by 2.3~3.3 (EG type). The third issue concerns ocean loading effect. A FG5 absolute gravimeter is used to collect time series of gravity value at a station within a network. The ocean loading effect is computed from the time series by least-squares harmonic analysis. The modeled loading is then used to correct for the loading effects of relative gravity measurements. Applying this approach to testing networks in Changhua and Yunlin Counties yields an improvement of gravity accuracy of 0.1~0.3 .