Search

Using GPS-levelling technique for height determination has been a very attractive method in recent years, due to rapid availability of high quality GPS positioning. With this method, a precise geoid is required to convert GPS ellipsoidal heights to orthometric heights. Conventionally, such a high quality local geoid is obtained through very dense gravity and GPS-levelling data (and other gravity related quantities, such as vertical deflections, gravity gradients). However, high quality and dense gravity and GPS-levelling data are not available for many developing countries, due to the cost and effort for such surveys. In this paper, a simple alternative method based on recent high quality global gravity and digital terrain models, which avoids the need of dense gravity and GPS-levelling data, is proposed. With this method, low and medium wavelength gravity field structures are estimated from the recent global gravity model, the Earth Gravitational Model 2008 (EGM2008), while the short wavelength structures are calculated from the latest global Digital Terrain Model (the Shuttle Radar Topography Mission (SRTM) elevation data). Residual topography reduced geoid undulation differences between GPS-levelling and EGM2008 are modeled as a trend and a corrective surface separately. As the GPS-levelling data used in this method is relatively sparse, an iterative cross-validation method is used to maximize the available data points in the corrective surface computation. As an example, a very rough region (in Canadian Rocky mountains) is selected to test this geoid determination method. The comparison of the geoid using the proposed method with the Canadian Gravimetric Geoid 2005 (CGG2005) in absolute and relative sense shows a slight improvement and the new geoid is able to be used to support GPS-levelling in the second and third order height networks. [ABSTRACT FROM AUTHOR]