Cadastral Data Migration: The 2.4 m Shift
You are the GIS manager for a utility company. Your team is migrating 12,000 km of cable records from a legacy CAD system to a new GIS. The legacy records are in HK1980 Grid. The new GIS uses WGS84/ITRF2014. A junior analyst used a free online converter with a single 7-parameter transformation and a fixed 2.4 m easting shift that he found in a blog. After loading 8,000 records, you notice that cables in Sai Kung appear offset by about 2.4 m relative to the government basemap, and some cables cross lot boundaries where they should not. The analyst says the offset is 'just a datum difference' and wants to finish the migration by Friday. A senior manager has already told the client that the data is 'survey grade'. Two field crews are about to use the new GIS to locate buried cables for excavation permits.
1. Diagnose the likely causes of the offset and explain why the analyst's approach is unsafe for this use.
Read the worked answer
The offset is unlikely to be a single fixed datum difference. HK1980 Grid coordinates and WGS84/ITRF2014 coordinates are related by a defined transformation, not by a constant 2.4 m easting shift. A free online converter may use unknown parameters, the wrong direction, the wrong ellipsoid or a different epoch. Applying a blog-derived shift to 8,000 records creates a systematic error that looks plausible on screen but is not survey grade.
The fact that cables cross lot boundaries is a warning sign. Cadastral boundaries and buried cable positions are both position-sensitive. A 2.4 m error can put a cable on the wrong side of a lot line, mislead a field crew and invalidate an excavation permit. The analyst's statement that it is 'just a datum difference' confuses a coordinate conversion with a verified transformation. The senior manager's 'survey grade' statement is also unsafe unless the data has been checked against control points and its accuracy is documented. The migration should stop until the transformation and QA are corrected.
2. What transformation and QA procedure should be used before the data is released?
Read the worked answer
I would specify the transformation from an authoritative source, not a blog. The procedure should state the source datum, target datum, ellipsoid, projection, transformation parameters, direction and epoch. If a 7-parameter transformation is used, it must be the published set for Hong Kong, and its residuals must be checked. If a grid-based transformation is more appropriate, that should be documented.
QA would include:
- a set of control points with known coordinates in both systems, distributed across the project area;
- computation of residuals at each check point, with mean, RMS and maximum;
- an independent field check at selected cable locations using a calibrated GNSS receiver or total station;
- a visual overlay against the government basemap and cadastral lot lines;
- a record of every transformation step, parameter set, software version and operator;
- a rollback plan and version control for the 8,000 loaded records.
The data should not be labelled 'survey grade' unless the accuracy is stated numerically and is fit for the intended use. For excavation permits, field crews should verify cable positions by locating them, not rely on the GIS alone.
3. What professional and ethical responsibilities arise from the manager's 'survey grade' statement and the field crews' use?
Read the worked answer
The first responsibility is to correct the record. If the data has not been verified to a stated accuracy, the client should be told that the 'survey grade' description is not supported. I would escalate the issue to the senior manager in writing, explain the 2.4 m systematic offset and the safety risk to the field crews, and recommend suspending release until QA is complete.
Second, I would ensure the field crews are notified immediately. They are about to use the GIS for excavation permits, where a positional error can damage buried services or injure people. They should be told to use the legacy records, survey the cables directly, or treat the GIS as indicative only.
Third, I would document the decisions, the data lineage, the transformation used and the limitations. I would not sign off the migration or describe it as survey grade to satisfy a Friday deadline. Professional competence and honesty require that uncertainty is disclosed, not hidden. If the manager insists on release, I would record my objection and seek advice through the appropriate professional channel.