EarthDefine Clutter3D
  • Telecom's most precise RF clutter dataset
  • Signal-aware classes, per-pixel heights
  • Nationwide at 1m resolution

Engineering-grade RF clutter data. Seamless nationwide coverage.

Industry-leading RF clutter intelligence

EarthDefine Clutter 3D is our top-of-the-line 20-class RF clutter dataset covering the entire continental US at 1-meter resolution. You get a rich 5-layer product with the same fidelity in rural Wyoming as in downtown Chicago, and no gaps. Every pixel carries an exact height value derived directly from LiDAR, to remove the need for estimations.

Plan with confidence, knowing that you are getting the highest value.


900×
finer than legacy 30m clutter data
>95%
classification accuracy
0-3 days
typical delivery time

See what's inside a Clutter 3D delivery

Toggle the layers below to explore all five datasets included in every order. If you'd prefer to work with the data directly, download a free sample and load it into your platform of choice.

Imagery DEM DSM Clutter 3D Buildings
Sample: Boulder, CO. See color strip below for clutter class schema.
10. Water
20. Bare Earth
31. Ground Veg
32. Low Veg
33. Tree, Low
34. Tree, High
41. Res. Low
42. Res. Med.
43. Res. High
51. Com. Low
52. Com. Med.
53. Com. High
61. Ind. Low
62. Ind. Med.
63. Ind. High
71. Parking
72. Roads
73. Railways
74. Airports
75. Other Imp.

20-class clutter schema

Most clutter datasets were designed for land use monitoring, not RF engineering. Class boundaries don't reflect how signal behaves, heights are estimated, and resolution is too coarse to distinguish a building from the tree line behind it.

Our 20-class schema is built differently: every class reflects RF propagation behavior, every pixel carries a LiDAR-measured height value, and the schema spans the full range of environments you'll encounter across the continental US. From open rural terrain to dense urban cores, no gaps. Explore the classes below to see exactly what you're working with.

Hover any row below to preview

Clutter ClassDescription
 10. Water Sea, ocean, river, pond, lake (excludes swimming pools)
 20. Bare Earth Soil, sand, rock, clearings
 30. Vegetation 4 classes Trees, shrubs, ground cover, crops
  31. Ground-level Vegetation Fields, grass, ground cover, low crops, turf
  32. Low Vegetation Shrubs and herbaceous
  33. Tree, Low Density Part of small clump
  34. Tree, High Density Part of large contiguous canopy
 40. Residential Buildings 3 classes Single/multi-family housing at varying densities
  41. Residential, Low Density Isolated structures (e.g., rural residential)
  42. Residential, Med. Density Buildings forming small clumps
  43. Residential, High Density Buildings forming large contiguous clumps
 50. Commercial Buildings 3 classes Office, mall, institution, community facility
  51. Commercial, Low Density Isolated buildings not part of clump, e.g., rural
  52. Commercial, Med. Density Buildings part of small clump, e.g., suburban corridor
  53. Commercial, High Density Buildings part of large clump, e.g., CBD
 60. Industrial Buildings 3 classes Factory, plant, heavy goods storage
  61. Industrial, Low Density Isolated industrial structures
  62. Industrial, Med. Density Buildings part of small clump
  63. Industrial, High Density Buildings part of large clump
 70. Transport & Impervious 5 classes Paved surfaces and transportation corridors
  71. Parking Surface parking lots and structures
  72. Roads All paved road surfaces
  73. Railways Rail corridors and associated structures
  74. Airports Runways, taxiways, and airport grounds
  75. Other Impervious Paved or hardened surfaces not classified elsewhere


20 classes. LiDAR heights. Building vectors. One stack.



Not all land cover data is built for RF planning

At 10m resolution, a tree and the building behind it can occupy the same pixel. Height values are estimated rather than measured. Your propagation model is only as accurate as your clutter data.

Here's how a typical 10m product compares to EarthDefine's 1m product for the same area.

ESA land cover
Hover to see basemap
European Space Agency (ESA) · 10m · 11 classes
EarthDefine Clutter 3D
Hover to see basemap
EarthDefine · 1m · 20 classes

EarthDefine's Clutter 3D is built to solve RF engineering problems: 1m resolution, LiDAR-measured heights, and 20 classes that reflect how signal actually interacts with the environment.



What This Means for Your RF Model

At 10m resolution, a tree canopy and the building behind it collapse into a single pixel, with an estimated height that reflects neither. Above 5 GHz, where almost any clutter feature acts as a high-loss obstacle, that simplification breaks down. In suburban and exurban terrain with moderate-to-dense tree cover, coarse clutter typically introduces 2–4 dB of error in Reference Signal Received Power (RSRP) prediction.

2–4 dB
typical RSRP error introduced by 10m clutter
+24%
RAN capacity gain per 1 dB modeling improvement
~0 dB
clutter-driven error with LiDAR-measured heights

Small improvements in clutter accuracy can materially change RF outcomes.

A 2–4 dB error is often the difference between predicted coverage and real-world performance.

EarthDefine's 1m clutter resolves obstructions individually. Trees, buildings, and ground features are individuated rather than averaged, and each class carries a LiDAR-measured height. The result is a propagation model that reflects what the signal actually encounters, not merely an averaged approximation of it.

Illustrative scenario over a 1.2 km C-band link in suburban tree-heavy terrain. Specific values may vary.

What 10m clutter tells your model Fresnel zone: miscalculated clearance "Shrubland": est. 4 m averaged · heights estimated 18 m · hidden in averaging subscriber (class unclear) predicted: clear, +14 dB margin · 18 m tree averaged into shrubland Bare / sparse veg Shrubland (single class · 18 m tree hidden inside) Built-up
European Space Agency (ESA) · 10m · 11 classes
What EarthDefine 1m tells your model First Fresnel zone: accurate clearance 1.5 m 1.2 m 18 m 2.0 m 1.0 m 1.8 m −6.2 dB loss subscriber (Residential, low density) predicted: marginal, +1.8 dB margin · pre-qualified before truck roll 20 32 32 33 32 32 32 41
EarthDefine · 1m · 20 classes · LiDAR-measured heights

In real deployments, the difference becomes immediately tangible:
• Links that actually perform in the field, not just on screen • Site candidates that can be confidently eliminated through accurate line-of-sight analysis • Installation outcomes that closely match prediction with fewer repeat truck rolls, and • Reliable decision-making even in dense, tree-heavy terrain where most propagation models begin to fail.

At this fidelity, you're removing one of the largest sources of error in RF planning.


Everything your RF model needs, in a single delivery


Pan-US coverage. Every market, every geography.

EarthDefine Clutter 3D is available as a seamless, ready-to-deliver dataset for any area within the United States. No stitching, no gaps, no waiting for data production.

3,125,942 square miles
Covering contiguous U.S., Puerto Rico, and the U.S. Virgin Islands.

50
States
3,235
Counties
32,629
Cities and places

Data coverage and freshness

NAIP acquires roughly half the continental US every year. As each new cycle completes, EarthDefine produces updated clutter data for affected areas, so your models always reflect the latest available imagery, not a legacy snapshot. Every dataset includes source imagery metadata to ensure that your predictions are reproducible and defensible.

EarthDefine Clutter 3D — US Coverage


Related Products

Tree Map

Tree Locations

Canopy Height Model (CHM)

Building Footprints 3D

Impervious Map