12.2 Vector Geometric Overlay: Intersect, Union, Identity, Erase, Symmetrical Difference & Update

12.2 Vector Geometric Overlay: Intersect, Union, Identity, Erase, Symmetrical Difference & Update

Quick Summary: Vector geometric overlay is the cornerstone of 2D spatial analysis. Unlike spatial selections or visual map displays, overlay operations execute computational geometry algorithms that calculate segment intersections, insert new topological nodes, subdivide features, and synthesize integrated attribute tables. The six foundational vector overlay operations—Intersect (Boolean AND), Union (Boolean OR), Identity (primary input plus overlapping identity features), Erase (Boolean NOT), Symmetrical Difference (Boolean XOR), and Update (geometric stamp/replacement)—follow strict mathematical rules governing supported geometry dimensions, spatial extents, and attribute inheritance. Managing post-overlay attribute validity (such as recalculating stale area/length values) and eliminating boundary-mismatch sliver polygons through cluster tolerance is critical for rigorous geospatial practice.

---

1. Computational Geometry Foundations of Vector Overlay

Vector overlay operates by integrating two distinct spatial layers: an Input Layer and an Overlay Layer (sometimes designated the identity, erase, or update layer). While simple spatial joins link tabular data based on geometric location without altering feature shapes, geometric overlay fundamentally reconstructs the spatial data model:

1. Segment Intersection Computation: The overlay engine applies line intersection algorithms (such as the Bentley-Ottmann sweep-line algorithm) to determine every coordinate location where an edge of the input layer intersects an edge of the overlay layer.
2. Node Creation & Line Splitting: At every intersection point, the engine inserts a new topological vertex (node) and splits the intersecting lines or polygon boundaries into distinct, non-crossing topological segments.
3. Planar Polygon Reconstruction: The newly subdivided boundary segments are linked to reconstruct closed planar polygons. Each newly created polygon represents a unique, homogeneous intersection of the input layers.
4. Attribute Schema Concatenation: The engine constructs an integrated attribute table containing columns from both input feature classes, populating each newly created polygon with the corresponding values of its source features.

   Input Layer (Soils)             Overlay Layer (Zoning)             Geometric Overlay Output
   +---------------+               +---------------+                  +-------+-------+
   |               |               |       |       |                  |  S1   |  S1   |
   |   Soil A      |       +       |  Z-1  |  Z-2  |       ====>      |  Z-1  |  Z-2  |
   |               |               |       |       |                  +-------+-------+
   +---------------+               +---------------+                  |  New Nodes Inserted  |
                                                                      |  Attributes Combined |

---

2. The Core Vector Overlay Operations

Vector overlay operations are governed by set theory and Boolean logic. Each tool produces distinct geometric configurations and attribute table propagation behaviors.

   INTERSECT (A AND B)        UNION (A OR B)             IDENTITY (A + (A ∩ B))
   +-----+                    +-----+-----+              +-----+
   |  A  |                    |  A  | A∩B |  B  |        |  A  | A∩B |
   |     +--+                 |     |     |     |        |     +-----+ (Only A extent
   |   |A∩B||                 +-----+-----+-----+        |     |       preserved)
   +---+----+                                            +-----+
   
   ERASE (A NOT B)            SYM. DIFFERENCE (A XOR B)  UPDATE (B replaces A)
   +-----+                    +-----+     +-----+        +-----+
   |  A  |                    |  A  |     |  B  |        |  A  |  B  | (B geometry &
   |     +                    |     |     |     |        |     +-----+  attributes
   |     | (B cut out)        +-----+     +-----+        |     |        overwrite A)
   +-----                     (Overlapping center removed)+-----+

1. Intersect (Boolean AND)

• Set Logic: $A \cap B$. Output features encompass only the geographic space common to all input datasets.
• Supported Input Geometries: Point, Line, or Polygon. Features can be overlaid across disparate dimensions (e.g., intersecting a line layer of roads with a polygon layer of flood zones).
• Output Geometry Dimension: The dimension of the output feature class is constrained to be less than or equal to the lowest dimension among the inputs. Intersecting two polygon layers can output polygons, lines (shared boundaries), or points (touching vertices). Intersecting a line layer with a polygon layer outputs lines or points, but never polygons.
• Attribute Behavior: The output attribute table combines all fields from all input datasets. Because output features exist solely where all inputs overlap, every field is fully populated without generating operational NULL values from non-overlapping areas.
• Primary Application: Identifying areas satisfying simultaneous criteria, such as locating timber stands that fall inside active mining claims.

2. Union (Boolean OR)

• Set Logic: $A \cup B$. Output features encompass the full spatial extent of all input layers combined.
• Supported Input Geometries: Polygon-on-Polygon ONLY! Points and lines cannot be unioned in standard vector geometric overlay.
• Output Geometry Dimension: Polygons (dimension 2).
• Attribute Behavior: Features in the overlapping zone ($A \cap B$) inherit populated attributes from both datasets. Features unique to Layer $A$ retain their $A$ attributes while all fields belonging to Layer $B$ are assigned NULL. Conversely, features unique to Layer $B$ retain their $B$ attributes while all fields from Layer $A$ are assigned NULL.
• Primary Application: Comprehensive land-use planning, landscape vulnerability assessments, and merging divergent thematic maps into an exhaustive polygon coverage.

3. Identity (Input Layer Priority)

• Set Logic: $A \cup (A \cap B)$. Computes the geometric intersection of the input features with the identity features. The output spatial extent is strictly identical to the spatial extent of the Input Layer.
• Supported Input Geometries: The Input Layer may be Point, Line, or Polygon. The Identity Layer must be a Polygon feature class (or share the same geometry dimension).
• Attribute Behavior: Input features that fall outside the boundary of the identity layer are preserved in their entirety; their identity attribute fields are populated with NULL. Where input features overlap the identity features, the identity attributes are joined.
• Primary Application: Enriching a primary dataset with administrative or environmental context (e.g., stamping county jurisdictional attributes onto a statewide parcel layer without truncating or expanding parcel boundaries).

4. Erase / Difference (Boolean NOT)

• Set Logic: $A \setminus B$ (or $A \text{ NOT } B$). Removes portions of the input features that overlap the erase features.
• Supported Input Geometries: The Input Layer can be Point, Line, or Polygon. The Erase Layer must be a Polygon feature class (or match the input dimension).
• Output Spatial Extent: Only the portions of the input layer falling outside the erase polygon boundaries are preserved.
• Attribute Behavior: The output dataset retains the exact attribute schema of the Input Layer only. No attribute fields from the erase layer are transferred to the output table.
• Primary Application: Removing exclusion zones, such as cutting water bodies, public rights-of-way, or steep slopes out of developable land inventory maps.

5. Symmetrical Difference (Boolean XOR)

• Set Logic: $(A \cup B) \setminus (A \cap B)$ (Exclusive OR). Retains all areas falling within either the input layer or the overlay layer, but strictly deletes the common overlapping areas.
• Supported Input Geometries: Polygon-on-Polygon (and in specialized engines, Line-on-Line).
• Attribute Behavior: Combines attribute columns from both layers. In areas originating from Layer $A$, Layer $B$ attributes are set to NULL; in areas originating from Layer $B$, Layer $A$ attributes are set to NULL.
• Primary Application: Temporal change detection, such as identifying areas of forest clearing or urban expansion between two decadal land cover surveys while omitting static, unchanged land.

6. Update (Cut and Paste Replacement)

• Set Logic: The Update Layer geometry and attributes overwrite and replace the underlying Input Layer features in areas of spatial coincidence.
• Supported Input Geometries: Polygon-on-Polygon.
• Attribute Behavior: Where the update layer overlaps the input layer, the input geometry is deleted and replaced with the update geometry, and the update attributes take precedence. Outside the update boundary, original input features and attributes remain intact.
• Primary Application: Patching an enterprise basemap with newly digitized subdivision plats without re-digitizing or rebuilding the surrounding regional dataset.

---

3. Comprehensive Vector Overlay Comparison Matrix

---

4. Attribute Table Propagation & The Stale Geometry Trap

When vector features are split during an overlay operation, their non-spatial attributes are copied into every newly created fragment. For categorical attributes (such as SOIL_TYPE = 'Clay' or ZONING = 'R-1'), this propagation is logically sound.

However, for quantitative geometric attributes (such as AREA, PERIMETER, LENGTH, or calculated densities), this creates a catastrophic analytical hazard known as the Stale Geometry Trap:

   Original Polygon A:           Overlay Operation Slices Polygon A:     Output Table Records:
   +-----------------------+     +-----------+-----------+               Poly A1: Area_Acres = 100.0 (STALE!)
   | Poly A                |     | Poly A1   | Poly A2   |               Poly A2: Area_Acres = 100.0 (STALE!)
   | Area_Acres = 100.0    | ==> |           |           |               
   |                       |     +-----------+-----------+               * Note: User fields DO NOT
   +-----------------------+      (Each fragment is 50 ac)               * automatically update in shapefiles!

Static vs. Dynamic Geometry Fields

1. Unmanaged Storage (Shapefiles, Raw Database Tables): User-defined numeric fields like ACRES, SQ_METERS, or MILES are static numbers stored in table columns. When a 100-acre parcel is bisected by an Intersect or Union tool into two 50-acre polygons, both output records retain ACRES = 100.0. If an analyst calculates summary statistics without updating the table, the computed land area will be double the true physical size.
2. Managed Geodatabases (PostGIS, Modern Geodatabases): Managed geodatabase feature classes maintain automated geometry-tracking fields (such as PostGIS ST_Area(geom) or geodatabase Shape_Area / Shape_Length). While system-maintained fields recalculate dynamically upon feature modification, any custom user-created attribute fields remain static.

[!IMPORTANT] Following any vector overlay operation, an analyst must explicitly re-calculate geometric fields (e.g., executing Calculate Geometry or calling ST_Area() / ST_Length()) before utilizing those fields in mathematical calculations, land valuations, or statistical reporting.

---

5. The Sliver Polygon Problem & Cluster Tolerance

One of the most persistent operational challenges in vector overlay is the generation of sliver polygons (also termed spurious polygons). Slivers are tiny, elongated, narrow polygons that appear along the shared boundaries of overlaid datasets.

   Boundary Layer 1:  -----------------*-------------------*-------------
                                        \                 /
   Boundary Layer 2:  -------------------*---------------*---------------
                                         [ Sliver Polygon ]
                              (Tiny area generated by non-coincident digitizing)

Etiology of Sliver Polygons

Sliver polygons are not genuine geographical features; they are digital artifacts caused by:

• Independent Digitizing: Two administrative agencies digitize the same physical boundary (such as a river centerline, ridgeline, or highway) from different aerial photo sources or at different scales.
• Positional Inaccuracy & Scale Mismatch: Overlaying a 1:24,000 USGS topographic quadrangle boundary with a 1:1,200 survey-grade cadastral boundary results in thousands of small, crossing boundary discrepancies.
• Conflation & Projection Divergence: Reprojecting datasets using different transformation parameters can introduce microscopic sub-millimeter coordinates offsets that cross back and forth repeatedly.

Prevention via Cluster Tolerance (XY Tolerance)

To prevent millions of microscopic slivers from corrupting overlay topology, vector processing engines implement a Cluster Tolerance (or XY Snapping Tolerance):

• Mechanism: Cluster tolerance defines a mathematical threshold distance. During the initial pass of the overlay algorithm, any vertices and line segments that fall within the cluster tolerance of one another are forcefully snapped together into a single, shared coordinate location.
• Calibration: The cluster tolerance should be configured based on the positional accuracy of the least accurate input layer (typically one order of magnitude smaller than the known data resolution).

[!CAUTION] The Danger of Over-Snapping: If the cluster tolerance is set too large, genuine spatial detail will be obliterated. Vertices from small narrow lots, cul-de-sacs, or winding stream meanders will collapse onto neighboring lines, resulting in inverted polygons, pinched geometries, and corrupted boundary topology.

Post-Processing Remediation: The Eliminate Tool

When slivers persist following an overlay operation, analysts employ automated filtering and elimination workflows:

1. Query Slivers: Isolate sliver polygons by executing an attribute query combining very small area with high perimeter-to-area ratios (thinness ratio): WHERE Shape_Area < 25.0 AND (Shape_Length / Shape_Area) > 1.5.
2. Eliminate: Run the Eliminate geoprocessing tool. Eliminate merges the selected sliver polygons into adjoining neighboring polygons based on one of two user-selected rules:
   • Longest Shared Border: Merges the sliver into the neighbor with which it shares the most boundary length (standard cartographic preference to preserve natural boundary flow).
   • Largest Area: Merges the sliver into the neighboring polygon having the largest overall surface area.

---

6. Practical Scenario: Environmental Impact Assessment

Scenario Context

A transportation agency is planning a new 4-lane bypass corridor. The environmental team must identify which sections of private parcels will be seized via eminent domain and what proportion of those acquired segments contain protected forested wetlands.

Overlay Workflow Execution

   Step 1: Intersect(RightOfWay_Corridor, Cadastral_Parcels) 
           --> Yields 'Acquired_Parcels' (Output matches Corridor extent;
               combines parcel owner IDs, zoning, and acquisition bounds).
               
   Step 2: Calculate Geometry on 'Acquired_Parcels' 
           --> Updates stale 'Acres' field to reflect the newly subdivided,
               severed parcel segments.
               
   Step 3: Identity(Acquired_Parcels, Regulated_Wetlands)
           --> Preserves all acquired parcel boundaries;
               appends wetland classification attributes where wetlands exist;
               populates wetland fields with NULL where parcels are upland.
               
   Step 4: Query and Eliminate Slivers
           --> Removes tiny edge fragments (< 1 sq meter) resulting from
               parcel-wetland boundary digitizing offsets.

By executing Intersect first, the analyst isolates only the condemned land while capturing landowner attributes. Re-calculating geometry ensures compensation is computed on the acquired acreage rather than the parent parcel. The subsequent Identity overlay appends environmental constraints without altering the condemned parcel boundaries.

---

7. Common Exam Traps & Pitfalls

[!CAUTION] Exam Trap 12.2.1: Applying Union to Non-Polygon Features. A recurring exam question asks which tool can be used to combine all features and attributes from a stream network layer (lines) and a soil survey layer (polygons). Multiple-choice options frequently list Union. This is impossible! In standard vector GIS, Union is strictly a polygon-on-polygon tool. Attempting to run Union on lines or points will throw an immediate execution error. The correct tool to find overlapping stream/soil segments is Intersect.

[!CAUTION] Exam Trap 12.2.2: Expecting Erase to Transfer Attributes. Candidates frequently assume that the Erase tool functions like Intersect or Union by appending attribute columns from both datasets. Erase does not transfer any attributes from the erase layer. It acts strictly as a cookie-cutter to remove geometry. The output feature class contains exclusively the fields present in the original Input Layer.

[!CAUTION] Exam Trap 12.2.3: Stale User-Defined Area Fields. When an exam question describes a post-overlay table where calculated acreage appears drastically inflated, the cause is almost universally the failure to recalculate user-defined area fields. In shapefiles and non-geodatabase tables, geometric fields do not update automatically when features are sliced by vector overlay tools.