Network RTK, Canopy, and an ALTA Relative Positional Tolerance
A 240-acre ALTA/NSPS survey of a wooded parcel in a river valley. Table A requires the 0.07 ft + 50 ppm relative positional tolerance. Roughly 30% of the tract is second-growth hardwood; the remainder is open pasture. Network RTK is available through a VRS service; the nearest reference station is 14 km away and the site is 6 km from the network's outer boundary. You observe five boundary points with 300 epochs each, all fixed, reported horizontal RMS 0.03-0.06 ft and vertical RMS 0.09-0.15 ft.
You check into three NGS-published monuments: two concrete posts with brass disks 1.8 miles apart agree with your derived positions within 0.03 ft horizontally and 0.05 ft vertically. The third, a 1934 disk with a recovery note from 1998, is 0.11 ft horizontally from your derived position and 0.14 ft in elevation. The property corners on the wooded knoll will not hold a fixed solution for more than a few minutes at a time. The client wants the survey in 10 days.
1. Does this RTK work satisfy 0.07 ft + 50 ppm, and how do you demonstrate that it does?
Read the worked answer
The tolerance is relative, not absolute. It applies to the positional relationship between any two points shown on the survey. For a 4,000-ft boundary-to-boundary relationship on a 240-acre tract, the allowance is 0.07 + (50/1,000,000 x 4,000) = 0.07 + 0.20 = 0.27 ft. Reported repeatability of 0.03-0.06 ft is well inside that, and the 14-km network baseline is not disqualifying by itself.
Demonstration is the issue. A single 300-epoch occupation proves precision, not accuracy. To rely on RTK you need:
- observations of the same points in at least two independent sessions with different satellite geometry and times of day;
- at least one independent check by a different method - a total station traverse or a static GNSS baseline - between points at the extremes of the survey;
- checks to three or more published or well-established monuments, with residuals recorded;
- a closed traverse or least-squares adjustment showing residuals within the tolerance.
On the plat, state the method, number of occupations, check results, and the relative positional tolerance achieved. If any point pair exceeds 0.07 + 50 ppm, reobserve it or tie it by total station. Also recognize that network RTK at 14 km depends on the network's modeled ionosphere and troposphere; set a maximum baseline in your scope and do not extrapolate beyond the network boundary. Vertical RMS of 0.09-0.15 ft is not survey-grade orthometric height and should not be used for vertical Table A items without a geoid model and level checks.
2. The 1934 monument is 0.11 ft horizontally and 0.14 ft vertically from your derived position. Do you reject it, hold it, or average the three?
Read the worked answer
You do neither reject nor average blindly. Two monuments 1.8 miles apart agreeing within 0.03 ft is a strong check; the third at 0.11 ft is a flag, not a failure. For a 2-mile baseline the ALTA allowance is 0.07 + (50 ppm x 10,560 ft) = 0.60 ft, so 0.11 ft is comfortably inside the standard. It still deserves investigation because it may indicate disturbance.
Steps:
- Recover the 1934 disk, read the stamping, and compare with the 1998 recovery note and any later recovery.
- Inspect for tilting, frost heave, or replacement - check the witness post and reference ties.
- Occupy it for a long static or multi-session RTK observation and compare with the published coordinate, remembering that the published value carries its own accuracy.
- Tie it by total station to one of the good monuments to remove GNSS from the comparison.
- If the 0.11 ft holds and the monument is sound, hold all three in a constrained adjustment and report residuals. If evidence shows disturbance, hold only the sound monuments and document the rejection.
The vertical discrepancy matters more than the horizontal. If a Table A item requires vertical data, use a current geoid model and verify with a level run between benchmarks; do not certify elevations from RTK heights alone. Averaging three monuments with one suspect is a way to hide a blunder, not resolve it.
3. The wooded knoll will not hold a fixed solution. What is your field strategy, given the 10-day schedule?
Read the worked answer
Canopy is the real risk. Under second-growth hardwood, fixed solutions come and go, and multipath with cycle slips produces plausible-looking bad positions. Options in order of preference:
- Build control in the open - two or more points with good geometry, by RTK or static - then run a closed total station traverse into the woods and radial-shoot the corners. This is the standard answer for ALTA work under canopy.
- If GNSS is used under canopy, occupy each point in multiple sessions on different days, require fixed solutions with low RMS and adequate satellite geometry, and repeat until independent sessions agree within about 0.05 ft.
- Use a total station resection or a short traverse from open control for any corner that will not hold a fixed solution.
Also observe at least one azimuth pair or check into a known bearing to catch orientation blunders, keep instrument and antenna height records, and re-check a point at the start and end of each session to detect drift.
The schedule is not a defense to a bad survey. If 10 days cannot support multiple sessions and an independent check, propose total station methods, or phase the work and issue the plat only after the check data exist. Tell the client in writing which method you used and what residual was achieved; a plat that quietly mixes single-occupation RTK points with traversed points is a liability. Retain raw observations and session logs, because the ALTA certification is only as good as the evidence behind it.