Assessing the Accuracy of GEOID03 and GEOID09 in Connecticut

Thomas H. Meyer
Associate Professor of Geodesy
Department of Natural Resources and the Environment University of Connecticut
Friday, Mar. 26, 2010, 3:00pm
Chase 130

Comparing published NAVD 88 Helmert orthometric heights of First-Order bench marks against GPS-determined orthometric heights showed that GEOID03 and GEOID09 perform at their reported accuracy in Connecticut. GPS-determined orthometric heights were determined by subtracting geoid undulations from ellipsoid heights obtained from a network least-squares adjustment of GPS occupations in 2007 and 2008. A total of 70 markers were occupied in these stability classes: 22 class A, 10 class B, 11 class C, 2 class D bench marks, and 25 temporary marks with transferred elevations. Adjusted ellipsoid heights were compared against OPUS as a check. We found that: the GPS-determined orthometric heights of stability class A markers and the transfers are statistically lower than their published values but just barely; stability class B, C and D markers are also statistically lower in a manner consistent with subsidence; GEOID09 does not exhibit a statistically significant residual trend across Connecticut; and GEOID09 out-performed GEOID03. A "correction surface" is not recommended in spite of the geoid models being statistically different than the NAVD 88 heights because the uncertainties involved dominate the discrepancies. Instead, it is recommended that the vertical control network be re-observed. We tested networks in which no phase-center variations were applied and in which all vectors (include "substandard" vectors) were included. We found that PCV cannot be ignored and that including "poor" vectors did not affect the results.


Thomas Meyer is an Associate Professor of Geodesy in the Department of Natural Resources and the Environment at the University of Connecticut.