Two Datums, One Old Culvert Benchmark, and a GNSS Height
A crew is establishing control for a 42-acre subdivision in Ford County. The base receiver sits on a temporary point whose coordinates come from a single 45-minute RTK observation; rover shots use 12-minute occupations, and the network reports 0.03 ft horizontal and 0.06 ft vertical. At the project's primary point the GNSS solution gives an ellipsoid height h = 285.40 m, and the current geoid model gives N = −29.85 m at that location. The nearest published NGS benchmark, 4.2 miles away, is NAVD88 = 1,033.02 ft. The project's design drainage grades were prepared in 1978 on NGVD29, and the county engineer's published note for this area states NAVD88 = NGVD29 − 0.86 ft. A 1978 project benchmark, described as a chiseled square on a culvert headwall, reads 1,033.91 ft (NGVD29). The crew chief intends to "just use the old benchmark, since it's on site," set design elevations from it, and label the plat "elevations NGVD29." The county requires the U.S. survey foot (1 m = 3.2808333 ft).
1. Compute the NAVD88 elevation implied by the GNSS data, compare it with the NGS benchmark, and explain what must happen before either value is adopted.
Read the worked answer
Convert the ellipsoid height. H = h − N = 285.40 − (−29.85) = 315.25 m. In U.S. survey feet: 315.25 × 3.2808333 = 1,034.28 ft.
Compare. 1,034.28 − 1,033.02 = 1.26 ft. That is far outside the claimed 0.06-ft vertical precision and far outside geoid-model uncertainty, which in this region is on the order of a few hundredths to a tenth of a meter (roughly 0.1–0.3 ft), not 0.38 m.
So something is wrong, and it is not automatically the benchmark. Candidate causes: an antenna or rod height entered incorrectly, a bad base coordinate propagating into every rover shot, a baseline long enough that the network solution is not control-grade, poor satellite geometry during the occupations, or a benchmark that has moved - a chiseled square on a culvert headwall is a classic candidate for settlement or replacement.
Required before adoption: a static or rapid-static session at the primary point, or a closed level loop from the NGS mark to the site. Then evaluate the loop misclosure and the benchmark's description and stability. Do not "fix" the GNSS value with a single 1.26-ft shift; a uniform shift hides whichever error actually occurred and will not hold across the 42 acres.
2. Convert the design grade of 1,036.50 ft NAVD88 to NGVD29 for the 1978 plans, and use the 1978 benchmark to test which elevation source is consistent.
Read the worked answer
Conversion. The given relation is NAVD88 = NGVD29 − 0.86, so NGVD29 = NAVD88 + 0.86. Design grade 1,036.50 (NAVD88) → 1,037.36 ft (NGVD29).
Check the 1978 benchmark. 1,033.91 (NGVD29) − 0.86 = 1,033.05 ft in NAVD88.
Interpretation. The 1978 culvert benchmark converts to 1,033.05, and the NGS mark reads 1,033.02 - agreement within 0.03 ft. Two independent sources, one of them published by NGS, are mutually consistent after the area conversion. The GNSS-derived 1,034.28 is the outlier by more than a foot.
Working conclusion. Hold NAVD88 ≈ 1,033.02–1,033.05 as project control, pending a level loop to confirm the culvert mark has not moved. The old benchmark is usable for the legacy plans, but only through the conversion - do not set design grades directly from a 1978 NGVD29 value in a NAVD88 project.
Caution on the conversion. The −0.86 ft note is an area relation, not a universal constant. For points outside this immediate area, use the official datum-conversion tool at the actual latitude and longitude rather than extrapolating the offset.
3. What should the plat's elevation note say, and what is the risk in the crew chief's plan to label everything NGVD29?
Read the worked answer
The crew chief's plan is a datum error. Labeling GNSS-derived or NAVD88 elevations as NGVD29 misstates every number on the plat by about 0.86 ft - enough to invert drainage grades, mis-set sanitary sewer inverts, and create ponding against a slab. The label is not cosmetic; it is the meaning of the numbers.
What the note must contain: the datum actually used (NAVD88), the specific benchmark and its published elevation, the method used to establish control (level loop or static GNSS, with the closure), the conversion applied to the 1978 plans including the +0.86 ft relation and its source, and the unit (U.S. survey foot).
Also state the legacy relationship explicitly, for example: "1978 drainage grades shown on the referenced plans are NGVD29; add 0.86 ft to convert to NAVD88 for this area." That single sentence prevents the next surveyor or contractor from mixing datums.
File practice. Record the source of every elevation in the project file, including which mark was held and why. If a future check shows the culvert mark moved, the file should let someone reconstruct exactly what was held and what changes - otherwise the entire subdivision has to be re-leveled to find out.