精密水準測量是採逐站向前推進之測量方式，每一測段往往需觀測數十餘次，導致系統性的誤差有可能大量累積，但由於其高精度的要求，使人們對以往所未注意之微小系統誤差會特別注意和關切。

水準測量中，儀器與轉點之沉陷現象往往都不受到重視，一方面是認為其量甚微而予以忽略，另一方面則假設使用規範中所規定的觀測模式即完全可消除此類誤差。但轉點沉陷的現象並非只於讀數時，當水準儀從單數站換至偶數站且開始觀測的時間中，欲轉換水準尺面之轉點亦會有輕微的沉陷，且誤差量會隨著測站數的增加而累積。

本研究主要目的在於不同觀點下分析TWVD2001水準資料，並比對觀測時無須使用水準尺鐵墊之高鐵水準資料，因軌道旁每30~40公尺左右即有一固定點，故轉點不會發生沉陷的現象，證實兩者資料的確具明顯差異。本研究根據系統誤差可能累積之特性來估計TWVD2001殘餘系統誤差量值，並推論造成此原因是換站時的轉點沉陷所致。最後，經轉點沉陷模式的改正，TWVD2001每公里測段往返閉合差之平均值從1.21mm降為0.01mm，改善幅度甚為明顯。

The way of precise leveling is taking every station forwards step by step. All sections have to observe several times often, and it will be generous that the systemic error is possible accumulated. Because of the high precision, people will be very deeply concerned to the small system error not noticing in the past.

At all times, people don’t pay much attention to the subsidence of a spirit level and turning point in leveling. On the one hand, few amount of errors is neglected, on the other, errors could be eliminated when using an order of leveling. Not only on reading, the rods which will change the other side have slight subsidence when the spirit level stands from the odd number to the even station and begins observing. In addition, the value will be accumulated systematically with the number of setups.

The study lies in different views analyze TWVD2001 materials which section discrepancies have seriously positive situation, and compare to the Taiwan high speed rail leveling data that needn't use the plates while observing. For this reason, the phenomenon that subsidence of turning point will not happen because it has fixed points per 30~40 meters by the track. We must verify that there are obvious differences in the two materials, and estimate TWVD2001 remaining systematic error correction parameters according to the reason that the error may be accumulated. Besides, we should also infer that remaining systematic error correction parameters caused by rods subsidence but not the movement of turning plate between two setups. In conclusion, through the correction of turning point subsidence. The average of normalized section discrepancies lower from 1.21mm to 0.01mm. It is improved by a wide margin.

1. 尹鐘奇，1976，實用大地測量學。國彰出版社，台中。
2. 王成機，1988，台灣地區水準測量折射誤差之分析。國立成功大學航空測量研究所碩士論文，台南。
3. 內政部，2001，一等水準測量作業規範，台北。
4. 李彥弘、劉啟清，1988，精密水準測量鐵墊滑動誤差之探討。第七屆測量學術及應用研討會論文集，G201-G211頁。
5. 於宗侍、魯林成，1982，測量平差基礎。測繪出版社，北京。
6. 胡明城、魯福，1993，現代大地測量學上冊。測繪出版社，北京。
7. 胡明城，2003，現代大地測量學的理論及其應用。測繪出版社，北京。
8. 張嘉強，1987，精密水準測量中重力與大氣折射之系統誤差研究。國立成功大學航空測量研究所碩士論文，102頁，台南。
9. 張坤樹，2001，一等一級水準網測量工作總報告。中興測量有限公司，台中。
10.陳國華，2004，整合TWVD2001水準及GPS資料改近台灣區域性大地水準面模式以應用於GPS高程測量。國立成功大學測量及空間資訊研究所博士論文，台南。
11.陳順宇，1997，統計學。華泰書局，台北。
12.曾清涼、劉啟清，1988，水準測量的折射誤差分析（I）（II）。國科會專題研究報告，NSC77-0410-E006-17，台南。
13.曾清涼、楊名、劉啟清、余致義、林宏麟，2001，一等一級水準網測量督導查核工作總報告書。國立成功大學，台南。
14.曾清涼、楊名、劉啟清、余致義、林宏麟，2003，一等二級水準網測量督導查核工作總報告書。國立成功大學，台南。
15.楊俊志、劉宗泉，2004，數字水準儀的測量原理及其檢定。測繪出版社，北京。
16.寧津生、陳軍、晁定波，2000，數字地球與測繪。清華大學出版社，北京，暨南大學出版社，廣州。
17.劉啟清，1986，精密水準測量之系統誤差。第五屆測量學術及應用研討會論文集，國立成功大學，F1-F25頁。
18.蘇哲民，2003，TWVD2001一等水準觀測資料之系統誤差分析。國立成功大學測量及空間資訊研究所碩士論文，台南。
19.蕭宇伸，2002，嚴密正高改正。國立交通大學土木工程研究所碩士論文，新竹。
20.Craymer, M. R., 1984, Data series analysis and systematic effects in leveling, Survey Science Technical Report No. 3, University of Toronto, Ontario, Canada.
21.Craymer, M. R. and Vanicek, P., 1983, Autocorrelation functions in the search for systematic errors in leveling, Manuscripta Geodaetica, Vol. 8, pp. 332-341.
22.Craymer, M. R. and Vanicek, P., 1985, An investigation of systematic errors in Canadian leveling lines, NAVD symposium’85, Rockville, Maryland, U.S.A., pp. 441-447.
23.Heiskanen, W. A. and H. Moritz, 1967, Physical Geodesy, W. H. Freeman and Co., San Francisco, CA, U.S.A.
24.Hirvonen, R. A., 1960, New theory of the gravimetric geodesy, Publications of the Isostatic Institute of the IAG, 32, Helsinki, Finland.
25.Kukkamäki, T. J., 1980, Error Affecting Levelling, NAD Symposium, Ottawa, CIS, pp. 1-10.
26.Pelzer, H., 1983, Systematic Instrumental Error in Precise Levelling, Dümmler’s Verlag, pp. 3-17.
27.Rapp, R. H. and N. Balasubramania, 1992, Aconceptual Formulation of a World Height System, Department of Geodetic Science and Surveying Report No. 421, The Ohio State University, Columbus.
28.Rappleye, H. S., 1984, Manual of leveling computation and adjustment, Supersedes Coast and Geodetic Survey Special Publication No. 240, NOS/NOAA, Rockville, Maryland, U.S.A.
29.Richard, J. H., 1985, Minimizing systematic errors in leveling, NAVD symposium’85, Rockville, Maryland, U.S.A., pp. 261-273.
30.Schomaker, M. C. and Berry, R. M., 1981, Geodetic leveling, NOS/NOAA, Rockville, Maryland, U.S.A.
31.Seeber, G., 1993, Satellite Geodesy：Foundations, Methods, and Applications, Walter de Gruyter, Berlin, Germany.
32.Torge, W., 1991, Geodesy, 3rd Ed., Walter de Gruyter, New York, U.S.A.
33.Vanicek, P., 1991, Vertical datum and NAVD88, Surveying and Land Information System, Vol. 51, No. 2, pp. 83-86.
34.Vanicek, P., Carrera, G. H. and Craymer, M. R., 1985, Corrections for systematic errors in the Canadian leveling network, DSS Contract No. OST 83-00053, Geodetic Survey of Canada, Ottawa, Canada.
35.Vanicek, P. and Krakiwsky, E. J., 1986, Geodesy : the concepts, North-Holland Press, Amsterdam, Netherlands.