International Journal of Engineering Inventions e-ISSN: 2278-7461, p-ISSN: 2319-6491 Volume 11, Issue 2 [Mar.-Apr. 2022] PP: 25-32 www.ijeijournal.com Page | 25 GIS-Based Geoid Refinement by GNSS/Levelling Data: A Case Study of Egypt Gomaa M. Dawod and Hoda F. Mohamed Geodesy Dept., Survey Research Institute, National Water Research Center, Giza, EGYPT Corresponding Author: Gomaa M. Dawod ABSTRACT: As an alternative to the extensive process of entirely re-developing a geoid model when new geodetic datasets are available, this paper proposes a simple, efficient, and fast approach for geoid refinement within a Geographic Information Systems (GIS) environment. It investigates several mathematical approaches for incorporating Global Navigation Satellite Systems (GNSS)/Levelling datasets. Such methods include 2- parametres, 4-parameters, and 7-parameters regression, Inverse Weighted Distance (IDW) method, and the krigging geostatistical method. Based on the available data, the recent SRI 2021 national geoid model has been refined using the five approaches with 220 new GNSS/Levelling data points. Based on available data and attained results, it has been realized that all investigated methods generate roughly the same accuracy level, and an improvement of almost 10% has been achieved. Such a small level of enhancement might be accredited to the non-homogeneous spatial distribution of the utilized datasets over the country. The final developed geoid model, named SRI 2022, has overall accuracy equals ± 0.14 m. It is recommended that in order to achieve a 1-5 centimeter accuracy of a geoid model in Egypt, updating/establishing of both GNSS and Levelling networks, with a good homogenous spatial distribution, is a must. --------------------------------------------------------------------------------------------------------------------------------------- Date of Submission: 11-04-2022 Date of Acceptance: 28-04-2022 --------------------------------------------------------------------------------------------------------------------------------------- I. INTRODUCTION Geoid modelling comprises an essential duty for geodesists worldwide, particularly with the rapid growth of utilizing the Global Navigation Satellite Systems (GNSS) technology. A geoid model plays the most significant role in converting the GNSS-based ellipsoidal heights to orthometric heights or elevations related to the Mean Sea Level (MSL) datum usually utilized in surveying, mapping, and civil engineering. Several national or regional geoid models have been investigated recently in quite a few countries such as Indonesia [1], Chile and Spain [2], and Vietnam [3]. Furthermore, other geoid models have been developed on a local basis within a country, such as the west desert in Egypt [4], and Jeddah city in Saudi Arabia [5]. The accuracy of geoid models differs significantly from one region to another based on the precision, number, and spatial distribution of the utilized datasets. A level of 1-cm has been achieved for a geoid model in Colorado state in the United States of America [6] and even 5-mm accuracy has been reported in Estonia [7]. On a national basis, many geoid models have been developed in Egypt even nationally such as Saadon et al. [8] or locally such as Elshewy et al. [9]. Refinement of a geoid model, a global or a national one, is a procedure that took place after its original creation by incorporating more new geodetic datasets. Such a process has been investigated by several researchers in the last decade. For example, Al-Kherayef et al. [10] have examined the addition of new observed GNSS/Levelling datasets to the Saudi geoid named KSA-Geoid17. Also, Pasuya et al. [11] have analyzed the refinement of the gravimetric geoid of Malaysia by incorporating terrestrial, marine, and airborne gravity datasets. Similarly, Wang et al. [12] proposed the enlargement of the Chinese geoid model by adding satellite altimetry levelling datasets. The mathematical and statistical methods of such geoid refinement comprise an essential research topic where several approaches have been proposed. Such models include, among others, the moving least squares approach [13], polynomial regression and Artificial Neural Network (ANN) approach [14], the 4-paratmeter removal [15], finite elements based bivariate [16], and minimum curvature surface [17]. Traditionally as far as new GNSS/Levelling datasets are available, a new geoid national model is developed using one of the geoid modelling packages such as the GravSoft scientific program. Instead, this paper proposes a simple and fast approach for refinement of an existing geoid model. Thus, based on one of the most-recent geoid models of Egypt, the current research examined the incorporation of more GNSS/Levelling datasets to improve its accuracy on a national scale. Such refinement procedure is carried out within a Geographic Information Systems (GIS) environment where several mathematical and statistical methods are applied and compared.