ISG

   International Service for the Geoid

   

Search in ISG Website



Links:










      

      

CoverNewton's Bulletin


Issue n. 4, April 2009 - External Quality Evaluation Reports of EGM08   [pp. 252-274]



Evaluation of the Earth Gravitational Model 2008 using GPS-leveling and gravity data in China


Jiancheng Li, Jinsheng Ning, Dingbo Chao, Weiping Jiang

Download file


Abstract

The evaluation of both PGM07A and EGM08 has been carried out in China using the observed data sets and high resolution regional geoid models including the observations of 652 GPS/Leveling points within Chinese national A/B-order GPS networks and 1160 GPS/Leveling points distributed over seven provincial regions, as well as the data sets of 183201 2'×2'gridded mean gravity anomalies collected from five provincial regions. In the evaluation of PGM07A and EGM08, the other recent released Earth Geopotential Models (EGMs) including EIG01C, EIG03C, EIG04C, GGM01C, GGM02C, GGM01C* and GGM02C* are also used for comparisons. The statistic results of the comparisonsshow that the RMS and Std.D of the differences between the quasi-geoid heights derived from PGM07A/EGM08 and their corresponding ones determined from the GPS/Leveling data of 652 A/B-order testing points are ±0.358m, ±0.338m for PGM07A, and ±0.284m, ±0.257m for EGM08 respectively, which are remarkably less than the RMS and Std.D corresponding to the other EGMs used in the comparisons. The RMS and Std.D of the differences between the quasi-geoid heights derived from PGM07A/EGM08 and their corresponding ones determined from the GPS/Leveling data of total 1160 points in seven provincial networks are in the range of ±0.09m ~±0.31m with a mean of ±0.19m for PGM07A, and ±0.07m ~±0.24m with a mean of ±0.16m for EGM08. For the other seven EGMs, however, the corresponding range of the statistics is ±0.26m ~±0.80m with a mean of ±0.44m. The comparisons of the quasi-geoid derived fromPGM07A and EGM08 as well as the other EGMs with those determined from the GPS/Leveling data of all 1812 points within both A/B-order GPS/Leveling networks and seven provincial onesare also made. The RMS and Std.D of the differences obtained from the comparisons are ±0.269m, ±0.243m for PGM07A, ±0.222m, ±0.186m for EGM08, respectively, and for the other seven EGMs, the corresponding statistics are in the range of ±0.375m ~±0.680m. The results of comparisons above show that the quasi-geoid derived from PGM07A and EGM08 can provide a better approximation to the quasi-geoids determined from Chinese GPS/Leveling data including both the data of national A/B-order networks and the data of provincial GPS/leveling networks than an approximation to the same quasi-geoids provided by the quasi-geoids derived from the other seven EGMs. According to all statistical results of comparisons between the EGMs-derived quasi-geoids and GPS/Leveling based ones, however, EGM08 is better than PGM07A as applied in Chinese mainland. The quasi-geoid heights derived from EGM08 are more accurate than those derived from PGM07A by a range of ±0.01m ~±0.07m with anaverage of ±0.03m in the RMS and Std.D of comparisons. In addition, the comparisons of gravityanomalies derived from PGM07A and EGM08 with those measured in five provincial regions are made, and the statistical results of the comparisons show that the RMS and Std.D of the differences between PGM07A/EGM08-derived gravity anomalies and those measured in the regions are within the range of ±13 ~29 mGal, with a mean of ±18 mGal. It also indicates that the gridded mean anomalies derived from EGM08 are more accurate than those derived from PGA07 by a range of ±1~4 mGal with a mean of ±2.7 mGal in the RMS and Std.D comparison. In the general, the quasi-geoid and gravity anomalies computed from PGM07A and EGM08 have significantly improved the representation of the Chinese local gravity field as compared with the other seven models used in the evaluation. However, our testing results show that for both EGM08 and PGM07, there are some large (or small) biases of the quasi-geoids which deviate from the general level of the biases in some regions.