The Parcel That Jumps 1.2 Meters

No error. The two numbers are in different reference frames.

• The plat is in ground feet. Record distances are what you would measure at the site with a tape or an EDM, near the site's elevation.
• The GIS is in grid (state plane) feet, which are projected onto the ellipsoid or the projection surface.
• The relationship is: grid = ground × combined factor, where combined factor = grid (scale) factor × elevation factor.

Computation: 1,200.00 ft × 1.000108 = 1,200.13 ft. That is exactly the analyst's number.

So the GIS is displaying a faithful conversion of your plat, and the plat is correct as a record document. The 0.13 ft is a systematic, predictable 108 ppm mapping effect that grows with distance - at 5,000 ft it would be 0.54 ft.

What you should still check on your side: that the combined factor you published is the value at the site (it varies across the zone), and that your plat states units, datum, and that distances are ground distances. That statement is what lets the analyst reproduce your arithmetic instead of filing a discrepancy report.

Most likely cause: the US survey foot / international foot unit definition. The two feet differ by 2 ppm (1 ftUS = 1.000002 international ft). Applied to a coordinate rather than a distance, the error scales with the magnitude of the coordinate, not the size of the parcel.

Test it: 1,950,000 ft × 0.000002 = 3.90 ft ≈ 1.19 m. That matches the reported jump.

The classic scenario: the layer is defined in the coordinate system with one foot definition (e.g., a "Foot_US" projection file) while the numbers were generated or exported by software that treated the unit as an international foot - or the reverse. The values are correct; the metadata is not.

Confirmation tests:

• Does the shift shrink toward the false easting (2,000,000 ft)? A unit-definition error scales with coordinate magnitude and nearly vanishes near the origin; a datum error does not.
• Compare against a known NSRS monument's published NC 3200 coordinate. A unit error will show as a few feet; a datum error as tens of meters.
• Read the .prj string and the metadata for the unit name, and check the software's feet conversion constant.

Rule out the datum. NAD27 to NAD83(2011) in this region is a shift on the order of tens of meters, not 1.2 m, so the 1998 NAD27 plat is not the explanation - although it must still be transformed before any overlay with your NAD83(2011) work.

Deliver data, not an endorsement of a shifted parcel.

1. Send a file in the county's native system - NC 3200, US survey feet, NAD83(2011) - with the projection file and a metadata note stating the combined factor at the site (1.000108) and that the plat's distances are ground distances. If you send a DXF or shapefile in grid coordinates, say so explicitly and give the factor used.
2. Refuse the rubber sheet, in writing if needed. A global translation or warp applied to the parcel fabric would move every parcel in the county to accommodate one 1.2 m metadata error, and it would silently move other parcels relative to their own monuments. The error is a projection/unit definition problem; fix the definition, not the geometry.
3. Keep the distinction clear. The fabric is a tax-mapping index built from plats and deeds; it is not evidence of boundary location and must not be advertised as such. Your plat locates monuments; the fabric approximates them.
4. Flag the 1998 plat separately. If that subdivision is also in the fabric, confirm whether it was transformed from NAD27 to NAD83(2011) and note the source of any transformation. That is a real discrepancy of a different magnitude and deserves its own check before it is fused into the same fabric as your work.