Letters to the Editor

To the Editor:

This letter is in response to Paul G. Weatherford’s article titled “GEOID09 in Middle Tennessee” (July 2011). In his article, Mr. Weatherford addressed a significant problem and traced the source of it back to its root causes.

He makes clear that his problem is not so much with the GEOID09 model as it is with the ellipsoidal heights used to create the model. This letter is not to raise objections with the article, but rather to amplify a few items touched upon by Mr. Weatherford.

The ellipsoid heights he referred to were refined in the NAD 83(NSRS2007) National Adjustment to reduce the systematic effects that resulted from previous adjustments. The 2007 adjustment significantly improved the overall national geoid model by reducing ellipsoid height inconsistencies that occurred between states. However, Jim Waters and several others from Tennessee brought to our attention problems that existed in the heights determined for the middle Tennessee region. In particular, there were decimeter-level differences between ellipsoid heights from CORS/OPUS solutions and the NAD 83(NSRS2007) values loaded in the NGS Integrated Database (IDB) (i.e., on the datasheets).

Data from the region is now being closely analyzed by the Observation and Analysis Division at NGS. The goals of this analysis are to learn what happened, find a means of improving it, and ensure that the lessons are applied everywhere—not just in middle Tennessee. NGS is especially interested in remedying this issue now, as we are preparing to conduct the even more comprehensive National Adjustment of 2011 (NA2011), and we wish to avoid duplicating the problems incurred in the 2007 adjustment. NA2011 will resolve a number of issues but will primarily serve to better align the passive control coordinates stored in the NGS database with the results from the new Multi-Year CORS Solution.

Hybrid geoid models, such as GEOID09, are a direct product of the data used to construct them. GEOID09 was created using a set of control stations in the NGS IDB as of 2009 that have both GPS-derived ellipsoid heights and leveled orthometric heights, collectively referred to as the GPS Bench Marks of 2009 (GPSBM2009). However, the NGS IDB is dynamic. When new information is available for a set of control points (from new surveys, better data-processing of existing surveys, etc.), the associated values for those points change, and the hybrid geoid no longer matches the data in the NGS IDB. NA2011 is an example of the source of such changes. Hybrid models are a bit like computers; they rapidly become obsolete.

Hence, GEOID09 will not be “fixed” or updated; instead, a new model will be developed. The next model, GEOID12, will be created based on the results of NA2011, and problems such as those seen in middle Tennessee should be prevented. Data used for GEOID12 (GPSBM2012) will also be evaluated by NGS state advisors and coordinators. If you know of, or encounter, any problem benchmarks, please bring those to the attention of your state advisor/coordinator. Ensuring the data is well scrubbed will allow the GEOID12 model to be best optimized.

Ultimately, NGS will move away from hybrid models, such as GEOID09 or GEOID12. These models are tied to an increasingly expensive and difficult-to-maintain network of passive control points for defining a vertical datum with known meter-level problems. NGS’ Gravity for the Redefinition of the American Vertical Datum (GRAV-D) Project is designed to develop a new vertical datum that is easier to use and maintain than one based on a network of leveled passive marks. GRAV-D will improve the determination of orthometric heights, particularly those used to determine flood plains and evacuation routes. As Tennessee and many other states have painfully realized, not only coastal states experience flooding. Backward compatibility will, of course, be maintained.

I would like to thank Mr. Weatherford for his clear and easy-to-understand presentation of the material in his article. Many of the issues I have discussed here may be found in greater detail on the NGS website.

Daniel R. Roman, Ph.D., Geodesist
Geoid Team Lead and GRAV-D 
principal investigator
Geosciences Research Division
NOAA’s National Geodetic Survey

» Back to our October 2011 Issue

Website design and hosting provided by 270net Technologies in Frederick, Maryland.