A need for height modernization.

Our modern society is becoming increasingly dependent on geographic data that is accurately and reliably spatially referenced.1 The ability to establish highly accurate horizontal positions on features has become much simpler in the last 10 years, thanks to a full constellation of GPS satellites, a significant drop in the price of GPS receivers coupled with a significant rise in their technology, the establishment of statewide High Accuracy Reference Networks (HARNs), and the proliferation of the national Continuously Operating Reference Station (CORS) network. The ability to establish highly accurate elevation data, however, is becoming more difficult. We have a need, more so than ever before, for updating the vertical control network throughout the country.

The National Spatial Reference System (NSRS), managed by the National Geodetic Survey (NGS), is a consistent national coordinate system accessible through precise coordinates published on physical monuments located throughout the country. Information on these monuments can be easily retrieved from the NGS website (www.ngs.noaa.gov). The search options are powerful, with options for a rectangular or radial search based on user supplied geographic coordinates, or an option for returning all monuments that fall within a specified county or quadrangle map. Both horizontal and vertical control points are provided at the site.<

There are multiple problems or weaknesses, however, with the vertical component of the NSRS. The direction of NGS has changed significantly over the last 15 to 20 years, mainly due to budget constraints. They no longer have the money or manpower to maintain a dense network of vertical control points. Further, many of the monuments within the HARN established in most states within the last 15 years do not have published orthometric heights. To complicate matters, many of the bench marks that have been established throughout the country in the last 50 years are placed in locations that are not suitable for direct GPS observations. A large percentage of bench marks are located under or around tree canopies, in the walls of buildings, or in urban canyons that block satellite reception.

Moreover, development has destroyed many bench marks over the years. Two lane roads have been widened to four lanes. New subdivisions and industrial developments have popped up in locations that were not anticipated when the control points were first established. And bridges and box culverts that once served as a stable location for control points have now been replaced. In many areas of the country, more than half the bench marks established by NGS and its predecessor agencies over the years have been destroyed.

Many of the marks that remain have elevations that are questionable. Subsidence in areas like New Orleans (see "Louisiana's Vanishing Act," Latest News, POB February 2004), Houston and parts of California has had a pronounced effect on the accuracy of elevations in those areas. An analysis of precise leveling data by the NGS in New Orleans between 1951 and 1965 revealed subsidence of 10 to 12 inches was common. Further analysis during 1985 to 1991 indicated annual subsidence in the range of 0.35 inches per year.2 Additionally, frost heave can be a major problem in the northern half of the country. Finally, post glacial rebound in areas like the Great Lakes region is taking its toll. This area is still rising, or rebounding, from the loss of weight from the large ice sheets that covered the area during the last ice age some 20,000 years ago. This rebound can be as much as 1 cm per year.

Wisconsin's early efforts in height modernization are very revealing. Of the 43 bench marks with published NAVD 88 heights leveled to in Phase I of the Wisconsin Height Modernization Program, only 19 (44%) could be constrained in the final NGS vertical adjustment. The other 24 (56%) had elevation changes ranging from -20.4 cm (-8.0 inches) to 19.2 cm (+7.5 inches) from the published NAVD 88 heights.3 These are major differences.

Significant effort in height modernization is needed. More HARN monuments should have precise orthometric heights established by differential leveling. New control points are needed to densify areas that have a low density of vertical control. Conventional leveling to CORS would be especially beneficial in establishing precise NAVD 88 elevations at these stations. Finally, additional conventional vertical surveys coupled with GPS observations in strategic locations could be used to improve the existing geoid model. This last step is very important to improving the accuracy of elevations gained from GPS surveys.

A Specific Example

One of the best examples of the need for more reliable vertical information comes from a low-altitude, high-accuracy mapping project that Photo Science completed for the Kentucky Transportation Cabinet three years ago. Designers needed very accurate elevation data on an eight-mile stretch of Interstate 64 in central Kentucky. The goal was to provide vertical mapping accuracies in the range of 0.03' for this stretch of interstate. Low-altitude photography captured from a helicopter was used for the project.

One of the first tasks to be accomplished was the establishment of precise elevations on the 194 targets that were required for the mapping. Obviously, very precise control would be required to meet the accuracy requirements. Research was conducted at the NGS website for first- and second-order bench marks located in the project vicinity. Six bench marks were found during this search. All were second order, class 0 marks that were set in 1960 during work associated with the design and construction of the interstate.

There was significant concern, however, as four of the bench marks were listed as Stability C, "May hold, but of type commonly subject to surface motion." The other two bench marks were listed as Stability D, "Mark of questionable or unknown stability." It did not take long to verify there were problems with the published elevations of more than one bench mark. The problem becomes more complex and requires more effort to solve when more than one bench mark does not match the others in the control network.

A digital level was used for all level runs. Generally, procedures for second-order accuracy were used during this project. The noted exception was that invar boards were not used. The results were very good; all loop closures exceeded the requirements for second order work.

There were several short level loops, which began and ended on the same bench mark. The closures for these level runs averaged 2.0 mm with a range of 0.1 to 3.5 mm. After discovering discrepancies in the published elevations, a long level line was conducted from a published bench mark near the western limits of the project to another published bench mark near the eastern limits. The elevation of these bench marks appeared to be stable in the early analysis. This level run checked the published elevations of these bench marks within 5 mm in the 9.4-mile run. One other bench mark located near the middle of the project checked within 3 mm in the level run. The other three published elevations differed from 2 cm (0.066') to 3.2 cm (0.105') from the elevations gained from the level runs.

Since the horizontal position was being established on the same targets using static GPS procedures, the elevation differences among the six NGS control points from GPS could be closely reviewed. Furthermore, these GPS observations introduced two new vertical control points into the network as two of the three HARN points used in the GPS observations had published orthometric heights. These differences provided additional verification of what was seen in the conventional levels. Three of the control points simply did not fit with the others, and the differences were substantial.

Benefits of Height Modernization

The benefits of height modernization are significant, not only to the surveying, mapping and engineering communities, but also to many governmental agencies and the private sector, including construction firms, developers and property owners. Height modernization should lead to a decrease in cost while improving the quality and accuracy of most surveying projects.

Height modernization programs are up and running in a limited number of states like North Carolina, Wisconsin and California. Other states are hoping to begin modernization efforts in the near future. Establishing accurate elevations at project sites can be a significant project cost under the existing system. Many times level loops several miles in length are required. Access to more vertical control points in a denser network provides obvious benefits. Knowing that the published elevations are reliable would be comforting. Finally, having access to better geoid models, which would allow more accurate GPS elevations, would also be significant.