A Spatial Analysis of GEOID03 and GEOID09 in Connecticut

Abstract

AbstractThe National Geodetic Survey (NGS) recommends using a hybrid geoid model to derive orthometric heights from ellipsoid heights. The accuracy of GEOID03 and GEOID09 were assessed independently in Connecticut. The present research analyses the spatial behavior of residuals derived from the comparison of differential levelled NAVD 88 orthometric heights and GPS-derived orthometric heights (using GEOID03 & GEOID09) at 72 benchmarks in Connecticut. Both geometrical and geostatistical analyses were performed on the residuals. A planar regression model indicates a weak spatial relation for residuals derived from GEOID03. This weakness was not noted in the analysis of residuals derived from GEOID09. Results of a four-parameter regression model does not indicate any need for a correction surface. A kriging surface was created with a fitted spherical semivariogram model and suggests GEOID09 captures more spatial variability of geoid undulation than GEOID03 in Connecticut.