Combination of Gravity Disturbances and Gravity Anomalies for Geoid Determination: A Case Study in Semarang City, Central Java, Indonesia

Conversion of geodetic height to orthometric height requires geoid to transform geometric elevation above ellipsoid into physical elevation above mean sea level. The need of accurate geoid is increasing as many leveling benchmarks have lost and deformed due to city development and natural activities. This paper presents geoid determination based on combination of gravity disturbances data and gravity anomalies data. Gravity disturbances were computed from 185 terrestrial gravity data. Gravity data were measured on March 2016 using Scintrex CG-5 gravimeter. All gravity stations coordinates were measured using rapid static method of GNSS to achieve sub-meter accuracy. Gravity anomalies data for improving the accuracy of the geoid models were measured by some government and private agencies using analogue gravimeters. It consisted of 10,149 data and covered whole Java island which was not less than 2 arc degree of latitude by 10 arc degree of longitude. Gravity disturbances of the city represented local gravity data, gravity anomalies of Java island represented regional data, while EGM2008 represented global gravity data. Gravity anomalies Java were converted to gravity disturbances data using geoid undulation of EGM2008 by simple free air reduction. The converted data were then shifted to local gravity data system. Gravimetric geoid were computed using Remove-Compute-Restore scheme and integral of Hotine based on combination of local and regional data. Gravimetric geoid was validated on 30 geometric geoid points measured by static method of GNSS and leveling. These validation points were distributed along 51 km of leveling line. Accuracy test showed that average deviation of gravimetric geoid to geometric geoid was -0.773 m while standard deviation of geoid was ±0.042 m. Conclusion of this research was that combination of gravity disturbances data and gravity anomalies data could achieve centimeter level accuracy. For future research, it was recommended to apply stochastic weighting to combine local and regional gravity data.