新一代台灣大地水準面模式---防災、監測、測繪之應用

Abstract

全球定位系統(GPS)已對傳統的測繪產生革命性的變化，GPS已廣泛使用於防災及環境監測，例如，地層下陷及邊坡滑動。許多GPS應用中，需利用大地水準面模式(geoid)將橢球高轉為正高。絕大部分的工程應用使用正高系統，而非橢球高系統。新的測繪科技光達(Lidar)先測得地面橢球高，再以geoid將之轉為地面正高。新的航測數值高程模式(DEM) 製作則先以GPS點為高控資料求橢球高，再以geoid求正高。美國及加拿大正考慮以geoid模式及GPS重新定義國家高程基準。Geoid日增的重要為本計畫的主要原動力，本計畫將收集現有的陸地，海洋及空載重力，加以除錯及求出系統誤差，並定出其解析度及隨機誤差。內政部5公尺解析度的DEM將用以求geoid的短波長效應。由重力衛星任務CHAMP、GRACE及GOCE求得之全球重力模式將為geoid長波長部分之用。傳統的去除-計算-回復法將於geoid計算。最小二乘配置法(LSC)將殘餘重力轉換為殘餘大地起伏。利用一等水準點上GPS觀測之大地起伏，此geoid模式之精度將被評估而求出位置相依的geoid誤差估值。利用純重力及觀測大地起伏，一混合的geoid可求得。此新的geoid模式將被推廣應用到防災及環境監測、GPS水準、航測及光達之DEM製作及高程聯結。

The Global Positioning System (GPS) have revolutionized the conventional surveying and mapping practice, and has been used extensively in hazard mitigation and environmental monitoring. Examples are land subsidence detection and landslide monitoring. For many GPS applications, it is necessary to transform GPS-derived ellipsoidal height to orthometric heights (OHs) with a geoid model. The OH system, instead of the ellipsoidal height system, is used in most engineering applications. The emerging new technology Lidar first determines ellipsoidal heights of the surface, which are then converted to OHs with a geoid model. A new trend in photogrammetric determination of elevations is to first employ GPS height control in the mapping of elevations and then obtain OHs with a geoid model. A geoid model, together with GPS, is now under consideration in USA and Canada to define a new national vertical datum. The importance of geoid prompts the need to develop a new geoid model for Taiwan. To this end, existing land, marine and airborne gravity anomalies will be collected and analyzed for data outliers and systematic errors. The spatial resolutions and data noises will be determined. New gravity data collected over 2008-2011 will be merged with the existing gravity data in an optimal manner. The 5-m digital elevation model (DEM) of the Ministry of the Interior will be used to model the short-wave length part of the new geoid. A latest global gravity model based on the data of the CHAMP, GRACE and GOCE satellite missions will be used as the long wavelength part of the geoid. For the geoid determination, a standard remove-computation-restore procedure will be employed. The conversion from residual gravity anomalies to residual geoidal heights are made by least-squares collocation (LSC). New techniques, largely based on spectral combination and modified Stokes’ kernels, will be tested to compete with the LSC geoid determination. The new geoid model will be evaluated using “observed” geoidal heights at Taiwan’s first-order leveling benchmarks, and location-dependent errors of the geoid model will be given. A hybrid geoid model is determined using observed and gravimetric geoidal heights. Promotion of the new geoid model in hazard mitigation and environmental monitoring, GPS leveling, Lidar DEM generation and vertical datum connection will be made.