Abstract

Improved Geoid Height Models from Supplemental Data on Bench Marks
Track: Education
Authors: Gerald Mader, Daniel Roman

This presentation follows previous talks that focused on outreach and education of surveyors. It is designed to update surveyors on the latest research related to the development of geoid height models as tools for datum transformation between NAD 83 and NAVD 88. All previous geoid height models have relied upon control data existant in the National Geodetic Survey Integrated Database (NGSIDB) to develop the conversion surface between the datums. In GEOID09, nearly 20,000 such points were available. However, about half of these points were located in only four of the lower 48 states. Most of the nearly 500,000 bench marks that exist in the NGSIDB have never been occupied with a GPS receiver. Hence, the bulk of the points are unused in determining the conversion surface. Given the disparate distribution of the few points that were occupied with GPS and the desire to supplement these in sparsely covered regions, alternative control data is desirable. The Online Positioning User Service Database (OPUS-DB) is where surveyors have the option of storing their observations for the use of others. Many of these marks were obtained on leveled bench marks. In November of 2010, there were about 422 points pulled from OPUS-DB with 285 representing new bench marks and providing supplemental control not previously available. These points are spread across the country and provide significant improvement in many regions, especially the sparsely covered western states. The errors resulting from interpolation over hundreds of kilometers can result in dm to multi-dm level errors in the resulting geoid height model. The OPUS-DB determined points then supplement the existing coverage from the NGSIDB. Many of these gaps were filled, and this reduced the interpolation error for those regions. More over, a campaign can be put in place to identify the sparse regions and likely candidate bench marks to target and fill these gaps. State Advisers/Coordinators and various state surveying groups have begun to organize efforts to collect GPS observations on the previously uncoccupied bench marks and store them in OPUS-DB. For example, the distribution of control data from the NGSIDB points in Arizona have been examined, and that State Adviser has made known regions that require supplemental information in the OPUS-DB. These points will be examined as they become available and a determination made as to whether to incorporate them into future geoid height models. Preliminary analysis does indicate that there is some slight inferiority to the quality of OPUS-DB data in that the apparent error signal (noise) is generally about double that of NGSIDB data. However, noisier data can be accounted for using least squares collocation - missing signal due to gaps cannot be easily overcome. Use of OPUS-DB to supplement coverage shows great promise as a means of readily collecting information without the need for following the Bluebbook, but doing so where it can provide the most improvement to future geoid height models for datum transformations.