Course · ← 0012 COGO · 0014 Spatial Networks → · engine layer

Lesson 0013 · ~15 minutes

Parcel Fabric

Esri's answer to everything lessons 0004 and 0012 raised: parcels as a connected fabric with recorded lineage, dimensions as first-class data, and coordinates that improve by least-squares rather than by dragging vertices.

The idea

A parcel layer is a pile of independent polygons. A parcel fabric is a connected network of parcels sharing points and lines, where each parcel carries its COGO dimensions and every change is attributed to a record — the legal document that caused it.

By capturing the record that created or retired a parcel, lineage can be tracked in both directions.1 That single sentence is the whole design: parcels do not change, they are retired by one record and created by another, and you can walk the chain forwards or backwards.

Where you have met this before

Lesson 0004 argued that parcels have lifespans, predecessors and successors, and that identifiers must never be reissued. The parcel fabric is that argument implemented in a product — with the addition that the record (the plan, the deed, the court order) is the thing that carries the change, rather than a bare timestamp. Model the same idea yourself and you have reinvented it; use it and you inherit twenty years of edge cases already handled.

The pieces

ElementWhat it is
RecordThe legal document driving a change. Every created or retired feature points at the record responsible — this is what makes lineage traversable both ways.1
Parcel typeParallel fabrics for different realities in the same space — tax parcels, ownership parcels, easements, subdivisions — each its own polygon and line pair.
Parcel linesBoundary lines carrying COGO dimensions from lesson 0012. The measurements, not just the shape.
PointsShared corner points. Parcels are connected because they use the same point, which is what lets an adjustment move a whole neighbourhood coherently.
Connection linesLines tying a parcel into the fabric without being a boundary — for example to a control point or a reference corner. Exportable per record alongside polygons, lines and points.1
SeedsA seed marks the interior of an area enclosed by lines, so a polygon can be built from the lines rather than drawn by hand.
AdjustmentLinesStores adjusted and statistical data for adjusted parcel and connection lines after a least-squares run.1

Least-squares adjustment

This is the part with no equivalent in ordinary GIS work, and the reason surveyors take the fabric seriously.

The fabric can be adjusted by least squares to evaluate and improve the spatial accuracy of parcel points: the adjustment uses the COGO dimensions on parcel lines together with control points to estimate best-fit coordinates for the fabric's points.2

Two adjustment types, for two different jobs: a free network adjustment evaluates the internal consistency of the measurements without pinning them to control; a weighted / constrained adjustment ties the network to control points to improve absolute accuracy.2

Read what that implies. You are not editing geometry — you are supplying measurements and control, and letting a statistical adjustment compute the coordinates. New control arrives, the adjustment reruns, and the whole neighbourhood improves at once while the legal dimensions stay untouched. Vertex-dragging cannot do this: it improves one parcel by breaking its agreement with its neighbours.

If you are not on Esri

There is no drop-in open-source parcel fabric. What you do have is the ability to build its ideas, which are not proprietary:

  1. Records as first-class rows Every geometry version cites the record that created or retired it. This is lesson 0004's lineage plus lesson 0007's source table, joined up.
  2. Store dimensions beside geometry Direction, distance, radius, arc length on boundary lines — lesson 0012. Cheap to store, impossible to recover later.
  3. Share corner points A points table referenced by lines, rather than coordinates duplicated per polygon. This is what makes a coherent adjustment possible at all.
  4. Adjust with a real library Least-squares network adjustment is a solved mathematical problem with open implementations. It is a genuine build, not a weekend — cost it honestly, or scope it out and say so.

Drill 1 — Fabric or not?

Would a parcel fabric earn its keep here?

Drill 2 — Recall


Your primary source

Esri, Parcel fabric data model — the schema: records, parcel types, lines, points, connection lines, seeds, AdjustmentLines:
pro.arcgis.com — Parcel fabric data model

Then Least-squares adjustments and the parcel fabric for free versus constrained, and Introduction to the parcel fabric for orientation.

Ask your teacher. Worth pushing: "what would a records table look like in my 0007 schema?", "which open least-squares libraries are usable?", "can a fabric and FFP data coexist in one system?" — that last one is the real design question for a country programme.