Skip to content
FastFuels Docs
Searching in-progress documentation for v2 (Beta). Switch to the stable version for production results.

Fetch a point cloud from USGS 3DEP

You are viewing in-progress documentation for v2 (Beta). Switch to the stable version for the current production release.

POST /domains/{domain_id}/pointclouds/3dep fetches public airborne lidar from the USGS 3D Elevation Program, clips it to your domain, reprojects it to the domain’s CRS, and stores it as a point cloud you can build on.

3DEP is airborne by definition, so the result is always type: als. There is no acquisition type to choose.

Prerequisites

Section titled “Prerequisites”

  1. An API key: my-api-key.

  2. A domain: your-domain-id.

  3. A coverage check. Every rejection below is visible in advance — see Check 3DEP lidar coverage.

The whole flow in one script:

Check coverage, pin the acquisition, fetch, and poll
"""Check coverage, fetch a 3DEP point cloud, and wait for it to finish."""


import fastfuels_sdk.v2 as ff


ff.set_api_key("my-api-key")


domain = ff.Domain.from_id("your-domain-id")


# 1. Pre-flight: is there lidar here, and will it fit the budget?
POINT_BUDGET = 200_000_000
coverage = ff.point_clouds.check_3dep_coverage(domain)
if not coverage.available or coverage.estimated_point_count > POINT_BUDGET:
    raise SystemExit("This domain cannot be fetched — check coverage first.")


# 2. Pin the acquisitions the pre-flight found, so the fetch is reproducible.
pinned = [dataset.name for dataset in coverage.datasets]
point_cloud = ff.point_clouds.create_point_cloud_from_3dep(
    domain,
    datasets=pinned,
    name="Blackfoot ALS (pinned)",
    description="Pinned to a single 3DEP acquisition so the fetch is reproducible.",
    tags=["blackfoot", "3dep", "pinned"],
)


# 3. Poll until the fetch lands. wait() raises JobFailedError on failure.
point_cloud.wait()


summary = point_cloud.summary
source = point_cloud.to_dict()["source"]
print(f"{summary.point_count:,} points at {summary.density:.1f} pts/m2")
print(f"Classes present: {summary.point_classes}")
print(f"Coverage: {source['coverage_fraction']:.1%}")

Step 1 — Create the point cloud

Section titled “Step 1 — Create the point cloud”

The minimal request is metadata only. The backend prefers a single acquisition covering the whole domain, and otherwise combines the fewest that fill it:

POST pointclouds/3dep — let the backend choose
curl -X 'POST' \
  'https://api-v2-prod-nyvjyh5ywa-uw.a.run.app/domains/your-domain-id/pointclouds/3dep' \
  -H 'accept: application/json' \
  -H 'api-key: my-api-key' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Blackfoot ALS",
  "description": "USGS 3DEP airborne lidar over the Blackfoot River domain.",
  "tags": ["blackfoot", "3dep"]
}'

Pin the acquisition instead

Section titled “Pin the acquisition instead”

Prefer this. Pass datasets with names from the coverage check, in priority order — where two overlap, the one listed first wins:

POST pointclouds/3dep — pinned
curl -X 'POST' \
  'https://api-v2-prod-nyvjyh5ywa-uw.a.run.app/domains/your-domain-id/pointclouds/3dep' \
  -H 'accept: application/json' \
  -H 'api-key: my-api-key' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Blackfoot ALS (pinned)",
  "description": "Pinned to a single 3DEP acquisition so the fetch is reproducible.",
  "tags": ["blackfoot", "3dep", "pinned"],
  "datasets": ["MT_Statewide_P3_4_B21"]
}'

Record the id: your-point-cloud-id.

Step 2 — Poll to completed

Section titled “Step 2 — Poll to completed”

The fetch runs in the background: pending → running → completed, or failed. Expect roughly a minute for a domain of a few hundred thousand square metres.

GET the point cloud
curl -X 'GET' \
  'https://api-v2-prod-nyvjyh5ywa-uw.a.run.app/domains/your-domain-id/pointclouds/your-point-cloud-id' \
  -H 'accept: application/json' \
  -H 'api-key: my-api-key'

georeference and summary are null until the status reaches completed — they do not fill in progressively, so check status before reading either.

Three things the completed resource tells you:

  • summary.point_count and summary.density — what actually arrived, which is not what the pre-flight predicted:
Two bar charts comparing the pre-flight estimate with the delivered cloud. Points: 7,835,003 estimated versus 10,496,309 delivered. Density: 14.8 estimated versus 19.8 points per square metre delivered.

The pre-flight estimate against the finished cloud for this domain. Both figures came in about a third higher than predicted, because the catalog’s density is averaged over an acquisition’s whole published extent — including the parts of it that hold no points.

  • summary.point_classes — the ASPRS classes present. Here that is [1, 2, 7, 9, 18, 20]: ground (2) is classified, but there are no vegetation classes (3, 4, 5) at all. That is normal — many acquisitions classify only ground and leave everything else in class 1.
  • source — what was used, what you asked for, the coverage fraction, and when the catalog was read.

When the request is rejected

Section titled “When the request is rejected”

All four rejections happen before anything is dispatched, so a doomed fetch never leaves a failed point cloud behind.

{
  "detail": "No USGS 3DEP lidar is available for this domain. Check coverage with GET /domains/{domain_id}/pointclouds/3dep/coverage."
}

Nothing to fetch. Check coverage first, or use a canopy source that does not need lidar.

Two more responses have no example here because they cannot be induced on demand:

  • 429 — a quota was exceeded. Point clouds have the tightest count limits in v2; see About quotas.
  • 503 — the USGS 3DEP catalog is temporarily unreachable. Retry shortly.

Next steps

Section titled “Next steps”

Common pitfalls

Section titled “Common pitfalls”
  • Fetching without pinning, then expecting to reproduce it. Record source.datasets from the completed cloud and pin those names next time.
  • Reading summary or georeference while the status is pending. Both are null until the job lands.
  • Assuming coverage_fraction: 1.0 means no holes. It is computed from catalog boundary polygons. Verify against the cloud’s own bounds and point_count.
  • Taking a two-acquisition mosaic without looking at the densities. The seam propagates into everything you derive. Pin the better survey if the difference matters.
  • Expecting vegetation classes. point_classes without 3, 4, 5 does not mean the cloud has no vegetation — it means the vendor did not label it. What matters downstream is whether class 2 is present, because that decides whether a derived CHM gets measured or inferred ground.