GIS Erase Tool: Remove Features Inside Another Layer

What the erase tool does

Erase is the opposite of Clip. You give it two layers. The first one is the input layer: the data you want to keep, for example land parcels, roads or buildings. The second one is the erase layer: polygons that mark areas you want to remove, for example flood zones, protected areas or a lake. The tool cuts the erase polygons out of the input. Everything outside them stays. Everything inside them is removed.

The output has the same geometry type as the input. A road that runs through a flood zone is cut at the edge of the zone, and only the dry pieces stay. A parcel that is half inside the zone keeps only its dry half. A bus stop inside the zone disappears. The attribute table of the input is copied to the output unchanged, so every parcel keeps its ID, owner and land use. The erase layer gives only its shape, and its attributes are not added. ArcGIS Pro (Erase and Pairwise Erase) and QGIS (Difference) follow the same rule.

I built this tool because "remove everything inside these areas" is one of the most common GIS questions, and it should not need a desktop GIS and an advanced licence. Drop a zipped Shapefile, GeoJSON, KML, KMZ, GPX or WKT file on the map, or draw your own shapes and erase rectangles. Pick the input and the erase layer, click Erase, and download the result. All the geometry math runs in your browser, so your files stay on your computer. The map starts without a basemap, which keeps it fast. You can turn on a street or satellite basemap at any time.

Erase, Clip, Intersect and Difference: which one do you need?

These four overlay tools are easy to mix up, because they all use one layer to cut another. The difference is which part they keep and which attributes they carry. This table puts them side by side for an input layer A and an overlay layer B.

A quick test: if you run Clip and Erase on the same input with the same overlay, the two outputs together give back the original input. Clip keeps the inside, Erase keeps the outside, and nothing is lost or counted twice.

Supported file formats

You can load several files at once, and you can mix formats. Each file becomes its own layer in the Layers panel. A layer can hold points, lines and polygons together, and each of them is erased with the right method.

Each file can be up to 50 MB. For very large national datasets, a desktop GIS is still the better choice. See the comparison further down this page.

Technical details

Here is exactly what the tool does with your data. If you know Erase in ArcGIS Pro or Difference in QGIS, most of this will look familiar.

Input geometry types

Point, MultiPoint, LineString, MultiLineString, Polygon and MultiPolygon. One layer can mix all of them.

Erase geometry types

Polygon and MultiPolygon, with or without holes. A layer without polygons cannot be picked as the erase layer.

Several erase polygons

All polygons of the erase layer are merged (dissolved) into one mask before erasing. Where erase polygons overlap, the area is still removed only once.

Holes in the erase layer

A hole in an erase polygon is not part of the erase area. Features inside the hole stay. For example, an island inside an erased lake is kept.

Attributes

Every attribute of the input feature is copied unchanged. Attributes of the erase layer are not added. As an option, the tool adds erase_status (outside or erased) and erase_area_m2, erase_length_m or erase_points for what is left.

Coordinate system

WGS84 longitude and latitude (EPSG:4326), the same as GeoJSON. A zipped Shapefile with a .prj file is converted to WGS84 when it loads.

Geometry model

Cutting is done on a flat plane in degrees, the same model that Turf.js and most web GIS libraries use. Areas and lengths in the results panel are measured on the sphere (geodesic), in square metres and metres.

Boundary rule

The edge of an erase polygon counts as part of the erase area. A point that lies exactly on the edge is erased, and so is a line piece that runs along the edge. This makes Erase the exact opposite of the Clip Tool on this site, which keeps those features.

Elevation (Z)

Existing Z values stay on their vertices. A new vertex created where a line is cut gets a Z value interpolated between its two neighbours.

Multipart output

If an erase cuts a feature into several pieces, the pieces stay together as one multipart feature (MultiLineString, MultiPolygon or MultiPoint), as in ArcGIS Pro. Tick "Split multipart results" to get one feature per piece.

Invalid geometry

If one input feature has broken geometry (for example a self-crossing polygon), that feature is skipped and counted, and the rest of the layer is still erased.

Privacy

Everything runs in your browser. No file and no geometry is sent to a server.

How to use the erase tool, step by step

Here is the whole workflow with the built-in sample data: a city layer with parcels, roads and points, and three flood-zone polygons to erase from it. Your own files work the same way.

1

Step 1: Open the tool

The map opens empty and without a basemap, so nothing loads that you do not need. The toolbar has Point, Line, Polygon and Erase Rectangle drawing tools, an Upload button and a Sample Data button.

2

Step 2: Add an input layer and an erase layer

Drag and drop your files onto the map, or click Upload. Here I clicked Sample Data. It adds a city layer with parcels, roads and points, and a flood-zone layer with a floodway, a retention basin and a spillway that overlap each other.

3

Step 3: Pick the layers and options

In the Erase panel, choose the input layer (the data to keep) and the erase layer (the areas to remove). Only layers with polygons appear in the erase list. Choose whether you want the extra erase_status, area and length fields, and whether multipart results should be split.

4

Step 4: Click Erase Layer

The result appears as a new layer in solid color, and the input layer is hidden. The flood zones are now empty gaps in the city. The results box shows how many features were untouched, partly erased or fully erased, and the area, length and points that are left and that were removed.

5

Step 5: Check the result with a basemap and the attributes

Tick "Show basemap" to see the result on a street or satellite map. Click any feature to open its attributes. The original fields are all there, plus the optional erase fields.

6

Step 6: Download the erased layer

In the Export panel, pick the layer and download it as GeoJSON, KML or CSV with a WKT column. The small arrow next to each layer in the Layers panel is a quick GeoJSON download.

How to erase features in a GIS

Every GIS has this tool, but the name changes. ArcGIS calls it Erase, QGIS calls it Difference, and PostGIS and GeoPandas call the function difference. Here is how to run it in this tool, in ArcGIS Pro, in ArcMap and in QGIS, and how to script it.

The same erase in code

If you remove the same zones every week, a script saves time. All these examples remove flood zones from a parcels layer.

import arcpy

arcpy.env.workspace = r"C:\data\city.gdb"

# ArcGIS Pro: Analysis toolbox > Pairwise Overlay > Pairwise Erase
arcpy.analysis.PairwiseErase("parcels", "flood_zones", "parcels_dry")

# ArcGIS Pro: Analysis toolbox > Overlay > Erase
arcpy.analysis.Erase("parcels", "flood_zones", "parcels_dry")

# Older ArcMap syntax (same tool, old name)
arcpy.Erase_analysis("parcels", "flood_zones", "parcels_dry")

# QGIS Python console: Vector > Geoprocessing Tools > Difference
processing.run("native:difference", {
    "INPUT": "C:/data/parcels.shp",
    "OVERLAY": "C:/data/flood_zones.shp",
    "OUTPUT": "C:/data/parcels_dry.gpkg",
})

# GeoPandas (works for points, lines and polygons)
import geopandas as gpd
parcels = gpd.read_file("parcels.shp")
flood = gpd.read_file("flood_zones.shp").to_crs(parcels.crs)

parcels_dry = parcels.copy()
parcels_dry.geometry = parcels.geometry.difference(flood.union_all())
parcels_dry = parcels_dry[~parcels_dry.is_empty]

-- PostGIS
SELECT p.*, ST_Difference(p.geom, f.geom) AS geom_dry
FROM parcels p
CROSS JOIN (SELECT ST_Union(geom) AS geom FROM flood_zones) f
WHERE NOT ST_IsEmpty(ST_Difference(p.geom, f.geom));

My take: for one erase on a normal-sized file, this tool is the fastest way from file to result. In ArcGIS Pro, reach for Pairwise Erase first. For jobs that repeat, write the script once and reuse it.

How the tool works, step by step

This is the path your data takes from the moment you click Erase Layer. Every step runs in your browser.

1. Read and clean the input. Every file is turned into GeoJSON features in WGS84. A GeometryCollection is split into separate features with the same attributes, so the engine only sees simple geometry types.
2. Dissolve the erase layer. All polygons of the erase layer are merged into one geometry, the mask M, with a polygon union. Holes stay holes. Overlapping erase polygons become one shape, so an area that two zones cover is removed only once.
3. Quick bounding box test. The tool compares the bounding box of each input feature with the bounding box of M. If the two boxes do not touch, the feature cannot lose anything, so it is copied to the output at once without any detailed math. This is the reverse of Clip, where such features are dropped.
4. Erase points. Each point is tested with a point-in-polygon test against M. Points strictly outside M are kept. Points inside or on the edge are removed. For a MultiPoint, only the outside points stay.
5. Erase lines. Each line segment is tested against every edge of M that is near it. The tool finds every place where the segment crosses an edge and cuts it there. Then it tests the middle of each small piece. If the middle is outside M, the piece is kept. Pieces that touch each other are joined again into one line.
6. Erase polygons. Each polygon is reduced by M with a sweep-line boolean difference (Martinez-Rueda). The result can have new holes where an erase polygon sat fully inside a parcel, or several parts where an erase polygon cut it in two.
7. Copy attributes and add fields. Each output feature gets a copy of all attributes from its input feature. If you asked for it, the tool adds erase_status and the area, length or point count that is left. If you ticked the multipart option, multipart results are split into single parts here.
8. Report and draw. The tool counts untouched, partly erased, fully erased and skipped features, adds up the area, length and points that are left and that were removed, and draws the result as a new layer. The input layer is hidden so you can see the result clearly.

The code for lines is shared with the Clip Tool on this site, with one switch turned the other way: Clip keeps the pieces whose middle is inside the mask, and Erase keeps the pieces whose middle is outside. That is why the two tools always fit together without gaps or overlaps.

The theory and math behind erasing

Erase is set difference applied to geometry. Here is the math, from the set rule to the algorithms and the measurements.

Erase(A, M) = { a \ M  :  a ∈ A  and  a \ M ≠ ∅ }

a \ M = { p ∈ a  :  p ∉ M }

M = M₁ ∪ M₂ ∪ … ∪ Mₖ     (all erase polygons, merged first)

a = (a ∩ M) ∪ (a \ M)        Clip part + Erase part
(a ∩ M) ∩ (a \ M) = ∅        no overlap

area(a) = area(a ∩ M) + area(a \ M)

Point P = (x, y). Edge from (xi, yi) to (xj, yj).

The edge is crossed when:
  (yi > y) ≠ (yj > y)
  and  x < xi + (y − yi) · (xj − xi) / (yj − yi)

P is inside  ⇔  the number of crossed edges is odd
Erase keeps P  ⇔  P is not inside and not on an edge

Road:  p = (−5, 5)   q = (15, 5)   length 20

Piece (t)      Middle x   Test          Erase keeps
0.00 – 0.25    −2.5       outside       yes → (−5, 5) – (0, 5)
0.25 – 0.45     2.0       inside        no
0.45 – 0.55     5.0       in the hole   yes → (4, 5) – (6, 5)
0.55 – 0.75     8.0       inside        no
0.75 – 1.00    12.5       outside       yes → (10, 5) – (15, 5)

Erase result: MultiLineString, 3 pieces, 5 + 2 + 5 = 12 units
Clip result:  MultiLineString, 2 pieces, 4 + 4     =  8 units
12 + 8 = 20  (the whole road)

Ring area (λ = longitude, φ = latitude, in radians):
A = (R² / 2) · | Σ (λ(i+1) − λ(i−1)) · sin φi |

Segment length (haversine):
h = sin²(Δφ / 2) + cos φ1 · cos φ2 · sin²(Δλ / 2)
d = 2R · asin(√h)

R = 6 371 008.8 m

Erase in this tool, ArcGIS Pro and QGIS compared

All three give the same geometry for normal data. They differ in cost, set-up and how much data they can handle. Here is a side-by-side view.

My recommendation: if your data lives in an ArcGIS geodatabase, use Pairwise Erase in ArcGIS Pro. It is fast, it works at every licence level, and arcpy lets you repeat the job. If you want a full desktop GIS at no cost, QGIS Difference does the same job and reads almost any format.

This tool wins when you need a result in a minute: someone sends you a zipped Shapefile of parcels and a KML of no-build zones, and you need the parcels that are left. You do not have to install anything, and nothing leaves your computer. It is also a clear way to learn how Erase behaves, because you see every cut on the map.

10 real-world use cases

Erase answers the question "what is left after we take these areas away?". These are the jobs I see most often.

1. Buildable land after constraints. A developer starts with a parcel layer and erases flood zones, wetlands, protected areas and setback buffers. What is left is the land that can actually be built on, with every parcel ID and owner still attached.
2. Flood risk screening. A planner erases the flood zone from a buildings or roads layer to get the assets that stay dry. Running Clip on the same data gives the assets at risk. Together the two outputs cover every asset once.
3. Land cover without water. An analyst erases lakes and rivers from land-cover or soil polygons, so area statistics count only land. Islands inside the lakes are holes in the water layer, so they are kept automatically.
4. Farm fields minus no-spray zones. An agronomist erases buffers around streams, wells and houses from the field boundaries. The result is the area that can be sprayed or fertilised, ready for the machine or the application plan.
5. Wind and solar site search. Energy planners erase buffers around homes, airports, roads and protected habitats from a study area. The polygons that remain are candidate sites, and their area shows the possible capacity.
6. Network outside an existing service area. A water or fibre company erases its current service area from the road network. The roads that remain show where new pipes or cable would be needed to reach the next customers.
7. Habitat outside development. An ecologist erases planned development footprints from a habitat map. The result shows how much habitat would remain, and each remaining patch keeps its habitat type and condition score.
8. Cleaning GPS tracks. A runner or researcher erases private areas, such as a home zone, from GPX tracks before sharing them. The rest of the track stays, with elevation kept on every point.
9. Sampling outside excluded areas. Field scientists erase roads, buildings and water from a study area before placing random sample points. Every point then lands in a place that can actually be surveyed.
10. Teaching overlay analysis. Teachers show Clip and Erase side by side with the same sample data. Students see that the two results fit together like puzzle pieces, which makes the idea of set difference easy to understand.

Frequently asked questions

Eighteen questions people ask about erasing in GIS, with direct answers.

Libraries used and credits

This tool is built on open source software. Many thanks to the people who write and maintain these projects.

• Leaflet 1.9 · BSD-2-Clause · Volodymyr Agafonkin and contributors
  Draws the interactive map, the layers, the drawing tools and the popups.
• Turf.js 7 (@turf/difference, @turf/union, @turf/area, @turf/boolean-point-in-polygon) · MIT · Turf authors
  Polygon difference for erasing polygons, union for dissolving the erase layer, geodesic area, and the point-in-polygon test.
• polyclip-ts · MIT · Luiz Felipe Machado Barboza, based on polygon-clipping by Mike Fogel
  The boolean polygon engine behind Turf's difference and union. It uses the Martinez-Rueda sweep-line algorithm with arbitrary-precision math.
• point-in-polygon-hao · MIT · Rowan Winsemius
  The fast point-in-polygon test that decides which points and line pieces fall inside the erase area.
• shpjs · MIT · Calvin Metcalf
  Reads zipped Esri Shapefiles (.shp, .shx, .dbf, .prj, .cpg) in the browser.
• proj4js · MIT · proj4js contributors
  Converts projected Shapefiles to WGS84 using their .prj file (used by shpjs).
• @tmcw/togeojson · BSD-2-Clause · Tom MacWright
  Converts KML and GPX files to GeoJSON.
• JSZip · MIT or GPL-3.0 · Stuart Knightley
  Opens KMZ archives to read the KML file inside.
• @terraformer/wkt · MIT · Terraformer contributors
  Reads WKT input and writes the WKT geometry column in the CSV export.
• OpenStreetMap, Protomaps (hosted by KONAKTIVE), CARTO, Esri World Imagery · ODbL and provider terms · OpenStreetMap contributors, Protomaps, KONAKTIVE, CARTO, Esri
  Basemap tiles and imagery, shown only when you turn the basemap on.

The line cutting step (splitting segments at boundary crossings and testing each piece) was written for this site and is shared with the Clip Tool. The erase engine is checked by automated tests for lines through holes, points on the edge and in holes, polygons that become a frame around an island, overlapping erase polygons, Z interpolation, and the rule that Clip plus Erase gives back the whole input.

Related geoprocessing tools