RTK GNSS Setting-Out in a Dense Urban Corridor
A geoinformatics engineer is asked to set out a 2.4 km trunk sewer in a dense part of Lahore. The contractor's survey team uses a single-base RTK GNSS system with a 35 W radio. The base is set over a pillar whose coordinates come from a 20-minute static occupation processed against a single CORS station 48 km away. During the morning the rover reports 'Fixed' almost continuously, but two independently measured checkpoints on a concrete benchmark differ from the design values by 0.07 m in plan and 0.11 m in height. Afternoon checks at the same points differ by only 0.02 m in plan but still 0.09 m in height. The contractor wants to proceed because the horizontal tolerance is 0.05 m and the sewer invert tolerance is 0.03 m. The design drawings give invert levels relative to mean sea level, while the RTK output is ellipsoidal height. There is no geoid model loaded in the controller. The engineer must decide whether to accept, reject, or conditionally accept the setting-out.
1. Diagnose the likely sources of the plan and height discrepancies, and explain why a fixed solution does not by itself prove accuracy.
Read the worked answer
A fixed ambiguity solution only means the receiver has resolved carrier-phase integers; it does not validate the base coordinates, the radio link, the satellite geometry, or the height datum. The morning plan discrepancy of 0.07 m improving to 0.02 m in the afternoon points to a base-coordinate, atmospheric, or radio problem rather than random rover noise. The base coordinate came from a 20-minute static session against a CORS station 48 km away: that is a long baseline for rapid static, and a single session gives no redundancy. If the CORS coordinate is in a different reference frame or epoch from the design, a plan shift appears. The persistent 0.09–0.11 m height difference is almost certainly ellipsoidal height versus mean sea level; without a geoid model the controller cannot output orthometric height. The sewer invert tolerance of 0.03 m is tighter than RTK vertical repeatability in this environment, so RTK alone cannot control invert levels.
2. Recommend a verification and correction procedure before any setting-out is accepted.
Read the worked answer
First, re-occupy the base at a known government benchmark or a second CORS-derived point and compute the baseline in the same frame as the design. Second, use at least three independent checkpoints distributed around the work, not two points near each other; observe each for 10–20 s in fixed mode and repeat under a different satellite geometry. Third, compare ellipsoidal heights with a levelled benchmark and either load a validated geoid model or derive a local height correction from at least three benchmarks; do not apply a single constant without checking. Fourth, set out plan by RTK but transfer invert levels by total station or level from a benchmark. Fifth, if the horizontal check remains above 0.05 m, reject the base coordinate and re-process the static session with a longer occupation and multiple CORS stations. Conditional acceptance may be possible for plan setting-out only after the horizontal check passes; invert levels should not be accepted until levelling is done.
3. State what the engineer should record and communicate if work proceeds conditionally.
Read the worked answer
Record the base coordinates and their source, the CORS station used, occupation times, PDOP and satellite count, checkpoint coordinates and residuals, the geoid model or local height correction applied, and the fact that invert levels were transferred by level rather than RTK. Issue a non-conformance note stating that RTK ellipsoidal heights are not accepted for invert control. Communicate in writing to the contractor that conditional acceptance covers plan setting-out only and does not cover sewer invert levels. Keep the checkpoint observations and processing reports as project records, because they are the evidence that the acceptance decision was based on verification rather than on the receiver's fixed status.